INTRODUCTION
Dwarf males, defined as males at least 50% smaller than females, have appeared in several different lineages of metazoans, both free-living and parasitic, often when females are sedentary and the sex ratio is male-biased (Vollrath, Reference Vollrath1998). Often, male dwarfism is accompanied by sexual parasitism (i.e. males being nourished by, and dependent on, the females). Examples include, but are not restricted to, polychaetes (Rouse et al. Reference Rouse, Goffredi and Vrijenhoek2004), several groups of crustaceans (Charnov, Reference Charnov and Southward1987; Raibaut and Trilles, Reference Raibaut and Trilles1993, Hirst and Kiørboe, Reference Hirst and Kiørboe2014), gastropods (Hoagland, Reference Hoagland1978), echiuroids (Leutert, Reference Leutert and Reinboth1975), cycliophorans (Neves et al. Reference Neves, Cunha, Funch, Kristensen and Wanninger2010) and even anglerfish (Regan, Reference Regan1925).
Parasitism has evolved independently several times in crustaceans. This has given rise to a wide variety of forms and life cycles, and, with them, of reproductive strategies. Many species display some sort of sexual dimorphism, including differences in size (usually, but not always, female-biased), morphology, behaviour and degree of association with the host (Raibaut and Trilles, Reference Raibaut and Trilles1993). These include numberless variations, from species with parasitic females and free-living males, such as in Ergasilidae (Kabata, Reference Kabata1979), to hyperparasitic dwarf males, though the most extreme examples probably are the cryptogonochoristic species, such as rhizocephalans (Høeg, Reference Høeg, Harrison and Humes1992) and some copepods (e.g. Nagasawa et al. Reference Nagasawa, Bresciani and Lützen1988). In these, originally mistaken for hermaphrodites, the mature male is reduced to a functional testis inside the female.
Members of the family Chondracanthidae Milne Edwards, 1840 also provide good examples of sexual dimorphism. This family of highly modified copepods, parasitic on marine demersal fishes, infects the oral and branchial cavities of their host (Kabata, Reference Kabata1959; Cavaleiro and Santos, Reference Cavaleiro and Santos2011), and sometimes other sheltered microhabitats, like the cloaca or the nasal cavity of the host (Boxshall & Halsey, Reference Boxshall and Halsey2004). They have been reported from wild as well as cultured fish (Tang et al. Reference Tang, Andrews and Cobcroft2007) and wild fish used as broodstock (Nagasawa, Reference Nagasawa2015). All species display evident sexual dimorphism, with hyperparasitic males permanently attached to nuptial organs on their several times larger female counterparts (Østergaard and Boxshall, Reference Østergaard and Boxshall2004; Østergaard et al. Reference Østergaard, Boxshall and Quicke2005), and females as large as 30 times the length of the male have been reported (Ho, Reference Ho1970). In the course of their adaptation to a parasitic life, females, on the one hand, have been transformed though the loss of external thoracic segmentation, the loss and simplification of swimming appendages and the development of larger body sizes, compatible with higher egg output; the body is thus simplified into a cephalosome, a trunk and a genito-abdominal region. Males, on the other hand, range from almost cyclopiform, such as in Juanettia Wilson C.B., 1921, to highly altered forms, with virtually no abdomen, such as in Medesicaste Krøyer, 1863 (Kabata, Reference Kabata1979).
Østergaard et al. (Reference Østergaard, Boxshall and Quicke2005) claim that, if the male chondracanthid is fed by the female, then smaller males are a smaller burden on the female, leaving more resources for egg production. Nevertheless, as some phenotypic variability in size is expectable, larger males may be present that represent a larger drain on the female's resources, and thus affect egg production. However, to our knowledge, no study has yet focused on the impact of male size on the reproductive output in copepod species with hyperparasitic males.
Kabata (Reference Kabata1959) studied the ecology of several species of Acanthochondria Oakley 1927, a chondracanthid genus, including host and site selection, infection levels and reproductive output. Among these species is Acanthochondria cornuta (O.F. Müller, 1776), a branchial parasite of several species of commercially important flatfishes, including the European flounder, Platichthys flesus (Linnaeus, 1758) and the European plaice, Pleuronectes platessa Linnaeus, 1758 (Kabata, Reference Kabata1959), often at high infection levels (Schmidt et al. Reference Schmidt, Zander, Körting and Steinhagen2003; Marques et al. Reference Marques, Teixeira and Cabral2006; Cavaleiro and Santos, Reference Cavaleiro and Santos2007, Reference Cavaleiro and Santos2009). This species shows a remarkable variability in morphology, ranging from short and flattened, to long and cylindrical, which led to lingering taxonomic debate until Ho (Reference Ho1970) provided conclusive evidence. The preferential infection site of the females changes as they grow, the water flow in the gills likely playing a major part in settlement (Cavaleiro and Santos, Reference Cavaleiro and Santos2011). Although Heegaard (Reference Heegaard1947) studied the larval development of A. cornuta, several stages are probably still undescribed, as he could not complete its cycle and probably missed several copepodid stages (Izawa, Reference Izawa1986). Otherwise, little of its biology is known.
The aim of this study is, therefore, several fold. Firstly, to gain a better understanding of the impact of dwarf, hyperparasitic males in the female's reproductive output. Secondly, to characterize the general reproductive strategy of Acanthochondria cornuta (O.F. Müller, 1776), and comparing it with what was reported for other groups of parasitic copepods. Finally, to compare the results with those reported by Kabata (Reference Kabata1959) for several Acanthochondria species, including A. cornuta.
MATERIAL AND METHODS
Sampling of parasites
Ovigerous females of A. cornuta were obtained from European flounder, Platichthys flesus, caught by demersal bottom trawling off Matosinhos, northern Portugal (41°10′N, 8°42′W). Collection took place once per season (i.e. spring, summer, autumn and winter) during 1 year, between September 2005 and May 2006 (for more information, see Cavaleiro and Santos, Reference Cavaleiro and Santos2009). The fish specimens were then frozen, until they could be examined for parasites under a stereomicroscope. Thirty ovigerous females were isolated from fish of each seasonal sample, in a total of 120 specimens. The isolated copepods were cleaned in 3·5% NaCl solution and fixed in 70% ethanol and later cleared for circa 1 h in 90% lactic acid, following Humes and Gooding (Reference Humes and Gooding1964). The four seasonal samples of parasites were then used to characterize the reproductive strategy of A. cornuta, here understood as a model chondracanthid. This species was chosen as model, owing to the high prevalence and intensity levels reported for P. flesus inhabiting the surveyed area (see Cavaleiro and Santos, Reference Cavaleiro and Santos2007, Reference Cavaleiro and Santos2009), i.e. as it is easy to obtain in the surveyed area.
Impact of sexual dimorphism on reproductive output
In order to evaluate (i) the relative contribution of each sex for reproductive output and (ii) a possible negative impact on reproductive effort associated with directing resources for male feeding, females of A. cornuta and attached hyperparasitic males were measured. Furthermore, females’ reproductive effort was characterized using a protocol similar to that described in Cavaleiro and Santos (Reference Cavaleiro and Santos2014) and Frade et al. (Reference Frade, Cavaleiro and Santos2016). Protocol adaptations were defined since the studied species produces multiseriate egg sacs and oblong to nearly spherical eggs. Measurements were conducted under a stereomicroscope coupled to a computer with version 4·6·3 of Axiovision digital image processing software (Carl Zeiss Microimaging Inc., Thornwood, NY, USA). The total length of both females and males was measured, in millimetres (mm), from the tip of the head to the tip of the genitoabdomen. In the case of bigamous females (i.e. females with two attached males), the mean length of the two males was used. Sexual dimorphism ratio, here understood as the ratio between female total length and male total length, was determined for each pair/trio. Length was used to determine size, rather than mass (e.g. Hirst and Kiørboe, Reference Hirst and Kiørboe2014), because the specimens might have lost mass during the clearing process, while length is unlikely to change during preservation due to the support offered by the cuticle.
The reproductive output was studied by determining fecundity (egg number), egg size (i.e. egg volume) and total reproductive effort (the product of fecundity and egg size). Mean egg volume was determined by measuring the length and width of 30 eggs within a given sac. Such values were then used to calculate the volume of an ellipsoid, as the eggs ranged from spherical to oblong, using the formula:
$${\rm Egg\; volume} = \; \displaystyle{4 \over 3}\pi. \displaystyle{{{\rm egg\; length}} \over 2}.\left( {\displaystyle{{{\rm egg\; width}} \over 2}} \right)^2 $$
The egg sacs were considered approximately cylindrical and, accordingly, egg sac volume was determined using the following formula:
$${\rm Egg\; sac\; volume} = \; \pi. {\rm egg\; sac\; length}.\left( {\displaystyle{{{\rm egg\; sac\; width}} \over 2}} \right)^2 $$
Fecundity was then estimated by dividing egg sac volume by egg volume. SPSS for Windows, version 22·0, was used for all statistical analyses, and results were considered significant for P < 0·05. Total length (for both sexes), sexual dimorphism ratio, fecundity, egg volume and total reproductive effort were characterized with respect to mean, standard deviation (s.d.), range interval (RI, minimum-maximum), coefficient of variation (CV), 95% confidence interval for population mean (CI) and distribution (one sample Kolmogorov-Smirnov's test testing for a normal distribution). In order to evaluate the relative contribution of each sex for reproductive output, correlation tests were conducted between female and male total length and sexual dimorphism ratio on one side, and fecundity, egg volume and total reproductive effort on the other. Pearson's correlation test was used for male total length-fecundity, and sexual dimorphism ratio-fecundity. The Spearman's correlation test was conducted for the remaining pairs of non-parametric variables, i.e. female total length-fecundity, female total length-egg volume, female total length-total reproductive effort, male total length-egg volume, male total length-total reproductive effort, sexual dimorphism ratio-egg volume, sexual dimorphism ratio-total reproductive effort. The latter test was also used between female total length and male total length.
A further analysis was conducted to assess the influence of male size. Females were divided in 5 classes depending on the size of the male attached to them: no male, males less than 0·80 mm long, males (0·80–0·85 mm) long, males (0·85–0·90 mm) long, and males larger than 0·90 mm. A non-parametric Kruskal–Wallis H test was then used to check for differences in fecundity, mean egg volume and total reproductive effort between the 5 classes, followed, when a significant difference was detected, by a Mann–Whitney U test between each pair of classes.
Seasonal variation
Box-and-whisker plots were used to depict the variation of all recorded variables across the four seasons. Also in this way, it was possible to get a general picture of the reproductive dynamics of A. cornuta.
Multiple sample comparisons, followed by pairwise comparisons, were also employed for seasonal samples. The tests used varied according to the characteristics of the data. A one-way analysis of variance test (one-way ANOVA) was used for multiple sample comparisons, for the variables that met the assumptions of normality and homoscedasticity, followed by pairwise comparisons through Tukey's test when the first returned significant results. In the case of variables that followed a normal distribution but did not have homogeneity of variances (based on the results of Levene's test), the Welch's robust test was used in multiple sample comparisons, followed by pairwise sample comparisons using Dunnett's T3 test. In the case of variables that did not follow a normal distribution, the Kruskal–Wallis’ test was used in multiple sample comparisons and the Mann–Whitney's U test was used in pairwise sample comparisons.
A multivariate analysis of variance (MANOVA) was run while conducting a discriminant function analysis (DFA; method: independent variables entered together), which permitted the evaluation of differences in body size and descriptors of reproductive effort between the females of A. cornuta in different seasonal samples.
Reproductive strategy
The skewness and kurtosis values for the distributions of fecundity and egg volume were determined and used to evaluate the general reproductive effort. The constraints of egg production were evaluated using a general linear model with type III sum of squares (general linear model (GLM) multivariate analysis). Specifications were as follows: fecundity and egg volume entered simultaneously as dependent variables; and season as a fixed factor; and female and male total length as covariates.
In order to evaluate the possibility of a trade-off between egg number and egg size, and position in the r/K spectrum of reproductive strategies, fecundity and egg volume data were plotted on the same graph, after arranging the specimens by ascending fecundity. To verify this, a non-parametric partial rank correlation test between fecundity and egg volume was then conducted for each seasonal sample. This test was complemented by observing the relative variation of fecundity and egg volume in the aforementioned boxplots.
RESULTS
The results of the measurements for the entire sample of A. cornuta and their egg sacs are presented in Table 1. Of the 120 females, 4 females (3·33%) had no male at the moment of sampling and only one (0·83%) was bigamous. The average female (6·635 mm) was 7·616 times longer than the average attached male (0·873 mm). Furthermore, female total length presented a wider variation (CV = 10·8%) than that of males (4·0%). While female length and reproductive output varied significantly across seasonal samples, male length did not. As a consequence, seasonal differences in sexual dimorphism ratio reflected variation in female body size rather than those of male body size (Fig. 1), and larger females did not hold necessarily larger males (i.e. no significant correlation was observed between the sizes of the two sexes).
Boxplots showing male and female total length, sexual dimorphism ratio, mean egg volume, fecundity and total reproductive effort for each of the four seasonal samples of ovigerous Acanthochondria cornuta females. In each box-and-whisker plot, the box represents the 25th percentile and the 75th percentile, the line in the centre represents the median, and the whiskers represent 1·5 times the interquartile range, while outliers are represented differently whether they are more than 1·5 times the interquartile range (circles), or more 3 times the interquartile range (asterisks).
Body dimensions and measures of reproductive effort (mean ± s.d. (range interval, RI), coefficient of variation (CV), limits of the 95% interval (CI), and results of the Kolmogorov–Smirnov test for normality evaluation) for the entire sample of ovigerous females of Acanthochondria cornuta
*Significant result (P < 0·05).
The results of the correlation tests are given in Table 2. Overall, sexual dimorphism ratio revealed statistically detectable correlations, of the same type (i.e. positive or negative), with the same variables as female total length, except egg volume, with which only sexual dimorphism ratio revealed a significant negative correlation (Spearman correlation coefficient = −0·209; P = 0·024). No statistically significant correlations involving male total length were observed. As female total length increased, so did fecundity (Spearman correlation coefficient = 0·281; P = 0·002) and total reproductive effort (Spearman correlation coefficient = 0·271; P = 0·003). Regarding the different male size classes, statistically detectable differences were observed only in fecundity (χ2 = 10·773, P = 0·029), and only between females from the class with the largest males (equal or over 90 mm long) and each of the two classes that preceded it, (80–85) mm and (85–90) mm. Females of the (80–85) and (85–90) classes had a median of 1080 and 1109 eggs, respectively, while the median for those of the >90 class was only 710 eggs (Mann–Whitney U = 205·0, n 1 = 24, n 2 = 28, P = 0·016 two-tailed, and Mann–Whitney U = 545·0, n 1 = 61, n 2 = 28, P = 0·006, respectively).
Results of correlation tests conducted for the entire sample of Acanthochondria cornuta females
*Significant result (P < 0·05).
Following the results of the Kolmogorov–Smirnov test, the distribution of female total length (P < 0·0001), egg volume (P < 0·0001), and total reproductive effort (P = 0·002) did not fit a normal distribution. The skewness for fecundity and egg volume was 0·221 and 1·015, while the kurtosis was −0·247 and 0·903.
The results of the comparisons between seasons are presented in Tables 3 and 4. Statistically detectable differences were observed among seasonal samples for all variables except male total length (Table 3, Fig. 1). Overall, summer females showed similar tendencies to autumn ones, and winter females to spring ones. On average, females were smallest in summer (6·213 mm) and autumn (6·435 mm) and longest in winter (7·159 mm). Spring females (6·721 mm) were significantly smaller than those of winter, but longer than those of summer. Very similar seasonal trends were observed for fecundity: summer and autumn females were similarly fecund (623 and 812 eggs, respectively), while winter and spring females displayed comparable high fecundity (1523 and 1361 eggs). Egg volume was highest in summer (1·57 × 10−3 mm3) and autumn (1·53 × 10−3 mm3), and lowest in winter (1·24 × 10−3 mm3) and spring (1·27 × 10−3 mm3). Total reproductive effort variation was more similar to those of fecundity than to egg volume. This apparent grouping of summer-autumn vs winter-spring females also accurately depicts differences in the reproductive strategy when it comes to an r/K strategy conceptual approach. Females of the former group produced a small quantity of comparatively large eggs, while those of the latter invested more on the quantity of eggs and less on egg volume.
Results of Levene's test and multiple sample comparisons of body dimensions and measures of reproductive effort for seasonal samples of ovigerous Acanthochondria cornuta
The specific test that was conducted on each variable is indicated in bold.
*Significant result (P < 0·05).
Pairwise comparisons of body dimension and measures of reproductive effort between seasonal samples of ovigerous Acanthochondria cornuta females
The specific test used for each variable is indicated in bold.
*Significant result (P < 0·05).
A moderate negative correlation was observed between fecundity and egg volume for autumn (r s = −0·415, P = 0·023) and winter females (r s = −0·424, P = 0·020). This is also apparent in Fig. 2, as the two variables seem to display somewhat opposite trends.
Variability in mean egg volume and fecundity of Acanthochondria cornuta for each seasonal sample, with specimens arranged by ascending fecundity.
Regarding the results of the GLM, no significant effects were found for season, or female or male total length, or any interaction terms.
DISCUSSION
Østergaard et al. (Reference Østergaard, Boxshall and Quicke2005), in their analysis of sexual size dimorphism in Chondracanthidae, found evidence of an allometric relationship between males and females. Moreover, while chondracanthid females are on average larger than those of free-living copepod species, males are smaller (Hirst and Kiørboe, Reference Hirst and Kiørboe2014). Our data partially corroborate the existence of independent regulation mechanisms of size for each sex. Males exhibit little variation in size compared with their mates, both in the total sample (Table 1) and the seasonal samples (Tables 3 and 4, Fig. 1). The sizes of both sexes were not correlated (Table 2), and larger females did not carry significantly larger males. Male dwarfism, according to Ghiselin (Reference Ghiselin1974), may be a side effect of rapid maturation, and can be achieved through a reduction of the number, or the duration, of the larval stages (Vollrath, Reference Vollrath1998). Since male chondracanthids attach to the female at the second copepodid stage (Heegaard, Reference Heegaard1947), different seasonal size variation may stem from a shorter growth period for males resulting in a more limited exposure to variation in macroenvironmental parameters, either favourable or not.
Chondracanthid males are probably nourished by the female (Østergaard, Reference Østergaard2004; Østergaard and Boxshall, Reference Østergaard and Boxshall2004). Thus, larger males could represent a larger burden on the female, and possibly on her reproductive effort. Accordingly, females belonging to the class with the largest males (>90 mm long) had significantly lower fecundity than those of the two preceding classes. However, considering only the seasonal variation, no apparent effect of male size was observed.
Boxshall (Reference Boxshall1974a , Reference Boxshall b ) and Frade et al. (Reference Frade, Cavaleiro and Santos2016) proposed a model of alternating generations with different body sizes and reproductive strategies for Lepeophtheirus pectoralis (O.F. Müller, 1776) (Caligidae), the latter using copepods from the same host sample as the present paper, and relatively similar results have been obtained in Lepeophtheirus salmonis (Krøyer, 1837) (Ritchie et al. Reference Ritchie, Mordue, Pike, Rae, Boxshall and Defaye1993). The results reported here for A. cornuta support the presence of a similar life history strategy in females of this species as well. For nearly all variables similar results were obtained for summer and autumn, and winter and spring. Summer and autumn females were smaller and less fecund, but produced larger eggs, while those from winter and spring were comparatively larger and produced a larger number of smaller eggs. Considering the seasonal changes in infection levels described in Cavaleiro and Santos (Reference Cavaleiro and Santos2009), summer females may originate from the eggs produced by the larger and more fecund winter and spring females, giving rise to the high infection levels observed during summer. Many of these summer generation females probably die before autumn, but the ones that survive may account for the similarities in size and reproductive output between summer and autumn, and the low infection levels during autumn. Eggs produced by summer females develop into overwintering females, which mature over winter and spring, becoming larger and more fecund.
Similar seasonal variation in infection levels has been reported for L. pectoralis and A. cornuta (Cavaleiro and Santos, Reference Cavaleiro and Santos2009). Both species have greatest epizootic potential during summer, with higher levels of infection (Cavaleiro and Santos, Reference Cavaleiro and Santos2009). The two species also display highest fecundity in winter, when infection levels are low (Frade et al. Reference Frade, Cavaleiro and Santos2016). In flounder aquaculture systems, prophylaxis aimed at preventing outbreaks of these two copepods should thus take place during this season, preventing the development of the resulting summer individuals. However, while different generations seem to overlap in L. pectoralis (Frade et al. Reference Frade, Cavaleiro and Santos2016), no evidence for that was found in A. cornuta. Alternatively, the overlapping period may not have been covered by the sampling.
Other than copepods, an analogous pattern has been reported for a freshwater leech, Piscicola geometra (Linnaeus, 1761) by Malecha (Reference Malecha1984), who suggests that higher reproductive output may be necessary to compensate the harsh winter period. Indeed, this strategy is likely to be found in more species of parasites, especially among short-lived species from temperate areas, whose individuals may not survive until the most suitable breeding season. As pointed out for L. pectoralis (Frade et al. Reference Frade, Cavaleiro and Santos2016), smaller, less fecund females may also stem from higher intraspecific competition, as they are associated with higher infection levels, but this does not explain the trade-off between egg number and egg size observed in A. cornuta (Fig. 2).
In other parasitic copepods from temperate areas, higher female length and fecundity are often associated with cooler temperatures (Ritchie et al. Reference Ritchie, Mordue, Pike, Rae, Boxshall and Defaye1993; Heuch et al. Reference Heuch, Nordhagen and Schram2000; Nordhagen et al. Reference Nordhagen, Heuch and Schram2000; Bravo et al. Reference Bravo, Erranz and Lagos2009; Cavaleiro and Santos, Reference Cavaleiro and Santos2014; Frade et al. Reference Frade, Cavaleiro and Santos2016), and, in Octopicola superba Humes, 1957 (Octopicolidae), this higher fecundity is achieved at the expense of egg size (Cavaleiro and Santos, Reference Cavaleiro and Santos2014). In the results reported here, A. cornuta females too were largest and most fecund in winter, but with smaller eggs, placing this species near the r-selected end of the r/K scale. In contrast, for L. pectoralis from the same sample (Frade et al. Reference Frade, Cavaleiro and Santos2016), and for Lernanthropus cynoscicola Timi & Etchegoin, 1996 (Lernanthropidae) (Timi et al. Reference Timi, Lanfranchi and Poulin2005), no such trade-off was found, possibly due to the amphidromous nature of their hosts, as seasonal migrations to lower salinity areas may render hosts unavailable to the infective stages for some time, selecting for larger, well-provisioned eggs (Frade et al. Reference Frade, Cavaleiro and Santos2016) and the high motility of fish hosts, which selects for higher fecundity (Gotto, Reference Gotto1962; Poulin, Reference Poulin1995). Together, these factors would prevent selection towards either end of the r/K scale (Frade et al. Reference Frade, Cavaleiro and Santos2016).
Kabata (Reference Kabata1959) described the reproductive output of several species of Acanthochondria, including two that are now thought to be just two types of A. cornuta (Ho, Reference Ho1970), Acanthochondria cornuta sensu strictu and Acanthochondria flurae (Krøyer, 1863). Significant differences were found between these two types, the flurae type having the largest eggs and the highest fecundity among the species in study, while the typical type had the lowest fecundity and second smallest eggs. The average values reported here fall in between those of the two types of A. cornuta reported by Kabata (Reference Kabata1959), but each of the two putative generations falls closer to one of the types. In average, summer and autumn females were more similar to the typical type and winter and spring females to the flurae type, though with smaller eggs. These data suggest that the two distinct generations may in part account for the differences associated with the two types of A. cornuta. Kabata (Reference Kabata1959) may have created an artificial distinction by relying excessively on host taxonomy, as he then considered “A. flurae” as mostly specific to Hippoglossoides platessoides (Fabricius, 1780), and the typical A. cornuta to Pleuronectes platessa Linnaeus, 1758. In L. salmonis, different host species produced significant differences in the reproductive output (Pert et al. Reference Pert, Mordue (Luntz), O'Shea and Bricknell2012), and this may also be the case in A. cornuta. The other two species studied by Kabata (Reference Kabata1959), Acanthochondria limandae (Krøyer, 1863) and A. clavata (Bassett-Smith, 1896), have intermediate values.
ACKNOWLEDGEMENT
The authors would like to thank the two reviewers for their constructive criticism.
FINANCIAL SUPPORT
The work was partially supported by the Structured Program of R&D&I INNOVMAR – Innovation and Sustainability in the Management and Exploitation of Marine Resources, reference NORTE-01-0145-FEDER-000035, namely within the Research Line INSEAFOOD Innovation and valorization of seafood products: meeting local challenges and opportunities, within the R&D Institution CIIMAR (Interdisciplinary Centre of Marine and Environmental Research), supported by the Northern Regional Operational Programme (NORTE2020), through the European Regional Development Fund (ERDF) (F Cavaleiro and MJ Santos).
