Sex dimorphism in European sea bass (Dicentrarchus labrax L.): New insights into sex-related growth patterns during very early life stages

The European sea bass (Dicentrarchus labrax) exhibits female-biased sexual size dimorphism (SSD) early in development. New tagging techniques provide the opportunity to monitor individual sex-related growth during the post-larval and juvenile stages. We produced an experimental population through artificial fertilization and followed a rearing-temperature protocol (~16°C from hatching to 112 days post-hatching, dph; ~20°C from 117 to 358 dph) targeting a roughly balanced sex ratio. The fish were tagged with microchips between 61 and 96 dph in five tagging trials of 50 fish each; individual standard length (SL) was recorded through repeated biometric measurements performed between 83 to 110 dph via image analyses. Body weight (BW) was modelled using the traits measured on the digital pictures (i.e. area, perimeter and volume). At 117 dph, the fish were tagged with microtags and regularly measured for SL and BW until 335 dph. The experiment ended at 358 dph with the sexing of the fish. The sex-ratio at the end of the experiment was significantly in favor of the females (65.6% vs. 34.4%). The females were significantly longer and heavier than the males from 103 dph (~30 mm SL, ~0.44 g BW) to 165 dph, but the modeling of the growth curves suggests that differences in size already existed at 83 dph. A significant difference in the daily growth coefficient (DGC) was observed only between 96 and 103 dph, suggesting a physiological or biological change occurring during this period. The female-biased SSD pattern in European sea bass is thus strongly influenced by very early growth differences between sexes, as already shown in previous studies, and in any case long before gonadal sex differentiation has been started, and thus probably before sex has been determined. This leads to the hypothesis that early growth might be a cause rather than a consequence of sex differentiation in sea bass.


30
The phenomenon of sexual size dimorphism (SDD) is common in animal species, and it is represented by the 31 differences in average body size of adult males and females [1]. Female-biased SDD is explained as a situation 32 where females are larger than males, while male-biased SDD is the reverse situation.

33
Male-biased SDD has been described in various teleost fish species: in different tilapiine strains, adult males

39
Female-biased SDD is also a characteristic of the European sea bass (Dicentrarchus labrax L.), one of the 40 major aquaculture species in the Mediterranean area. The females of this species are known to be about 30% 41 heavier than the males from 300-400 g until over 1000 g [12,13]. Furthermore, the common aquaculture 42 practice of size grading has shown that the largest fish selected at 86 days post-hatching (dph) later result to 43 be mostly females [14]. Previous studies exploiting individual tagging suggested that females are already 44 significantly heavier than males from 105 dph (1024 degree days above 10 °C), with a stable 40% difference 45 from 197 to 289 dph [15].

46
This supports the hypothesis that in European sea bass sex-specific growth may happen before gonadal sex differentiation, as differentiation starts only around 128 dph [14]. In this species, there are no sex chromosomes through microchip ID reading, then placed over a light and prevent infections) and they were allowed to rest for 1 to 2 h before being returned to their rearing tank.

97
Fish were reared in a common garden tank from 117 to 358 dph with the same conditions described above

99
During the biometric measurement performed between 137 and 335 dph, fish were anesthetized as described 100 above, adjusting the anesthetic solution of ethyl-p-aminobenzoate and seawater according to the increasing 101 size of the fish, the microtag was read, the body weight and the standard length were individually registered. the sex was determined macroscopically through the direct observation of the gonads or using a gonadal squash

105
Prediction of body weight from digital picture measurements and prediction of standard length and sacrificed with an excess of anesthetic (MS-222) to directly measure the length and the weight of each fish 111 (total number of fish = 250). The standard length was obtained with a V-12B 12" vertical optical comparator 112 (Nikon) that allowed an accurate measure through magnification of the larva. The measure of the body weight 113 was achieved using a precision scale (to the nearest 0.01 g) after drying the fish with absorbent paper.

114
In addition, a digital picture of each fish was taken following the same procedure used for the experimental

117
A volume index was calculated for each fish from height and length as: Pearson's coefficient of correlation (r 2 ) between measurements obtained from image analysis and 120 measurements obtained directly was estimated in R using cor.test function (package stats, R version 3.5.0,

121
[23]). Multiple regression models using length, height, perimeter, area and volume were tested using lm and

127
The r 2 between estimated and measured BW as estimated to assess the accuracy of the prediction model.

128
During the biometric measurement performed at 117 dph, only BW was directly measured on the fish; for this 129 reason, a model to estimate standard length using body weight was built. The data from the 50 additional fish 130 sacrificed during each biometric measurement was used; standard length and body weight were log-131 transformed. The procedures followed were the same as the model built for body weight.

132
The predictive models were then applied to the experimental fish dataset to estimate the body weight of fish measurements. The formula was the following: where BW f is the final body weight, BW i is the initial body weight and t is the number of days.

Statistical analyses
140 The number of males and females in the population were compared through χ 2 tests.

149
Prediction of body weight from digital picture measurements and prediction of standard length 150 Pearson's coefficient of correlation (r 2 ) between measurements obtained from image analysis and 151 measurements obtained directly were all high and significant, ranging from 0.9533 to 0.9963 (see table S1 in

152
Supplementary material 3 for details). The traits with the greatest correlation with BW were area (0.9898, p <

162
The coefficient of correlation between the logarithm of the measured SL and the logarithm of the measured 163 BW was significantly high (0.9873, p < .0001). The logarithm of SL was estimated for the fish aged 117 dph 164 (when only BW was directly measured) with the following model, having an AIC equal to -1111.5 and an R 2 165 equal to 0.9708: The global r 2 of the regression between measured and estimated SL using model (2) (Table 1 and Table 2), which was significantly different (χ 2 176 = 13.364, p-value = 3 × 10 -4 ).

177
Differences in terms of growth patterns were observed between females and males (Table 1). On average, 178 females were longer compared to males from 103 dph, when females were 6% longer than males, and a 179 significant difference was maintained until 165 dph, with females close to 4% longer than males. From 180 at 103 dph (females were 20% heavier than males), between 117 and 165 dph (females were about 10% heavier than males), and at 265 dph. From 180 dph until the end of the experiment, the difference in weight between During the first three biometric measurements (at 83, 89 and 96 dph), even though the differences were not 188 significant, females were already around 2 to 6% longer and 13% heavier than males.

189
The daily growth coefficient (DGC) was higher in females in almost all the periods analyzed, with the only 190 exception of the interval between 103 and 110 dph (Table 3). Significant differences between males and 191 females were detected only during the interval between 96 and 103 dph, where the DGC of females was 32.5% 192 higher than that of males.
193 pers. comm, vs. 112 dph in our study). We were thus able to individually follow the growth of future males 226 and females starting from 0.16 g instead of their 0.59 g.

227
At 83 dph, the fish were 23 mm SL and 0.16 g BW, and males and females were not yet statistically different 228 in size. Then, SSD built up and from 103 dph (645 degree days above 10 °C) the differences between males 229 and females became significant, until 165 dph (1241 degree days above 10 °C). The time when SSD builds up 230 is also shown by the difference in growth rate, measured as DGC, which was 32% higher in females than in 231 males from 96 to 103 dph. We cannot completely rule out that SSD existed before this time, as females were

232
(not significantly) larger than males as of 83 dph, and the lack of significance may be caused by the limited 233 samples size at those ages. Indeed, the sample size during the first period (83 to 110 dph) was rather low, due 234 to the fact that fish were not all tagged at the beginning of the experiment but at different ages, as we did not 287 Funding

288
The study was funded by the French Ministry of Agriculture (CRECHE2019 project).

290
We wish to thank Intellibio (Seichamps, France) for providing technical support and instrumentation.

292
The authors declare that they have no competing interests.