DISCUSSION
Our study presents for the first time a detailed and an accurate picture of Chthamalus stellatus population distribution in the Mediterranean Sea. Previous studies (Crisp, Southward, & Southward, 1982; Pannacciulli et al.,1997; Shemesh et al., 2009) used only a single or a very few individuals to represent each local population. Using 63 SNPs in 14 populations, each consisting of 15–28 specimens, enabled us to get a more reliable picture of this barnacle’s populations in the Mediterranean as well as the Eastern Atlantic.
The various analyses performed in our study clearly demonstrate four distinct clusters (see fig. 2-5). The Eastern Atlantic (EA) cluster, consists of the western peripheral populations of C. stellatus , namely, Biarritz, the Canaries and Madeira. This cluster was long-established in previous studies that show the separation between the Atlantic and the Mediterranean populations of C. stellatus(Crisp et al., 1982; Pannacciulli et al.,1997; Shemesh et al., 2009). The Western Mediterranean (WM) cluster, which consists of Bizerte, Málaga and Melilla, is a sister cluster to the EA cluster. Two of the cluster’s locations, Málaga and Melilla, are geographically adjacent to the Strait of Gibraltar, and therefore naturally influenced by the surface current entering from the Atlantic and flowing along the Northern Africa coast (Millot & Taupier-Letage, 2005). The third location (Bizerte) is geographically quite distant, closer to the Mid-Mediterranean locations studied here, but probably still influenced by the flow along the coast. The resemblance between the Atlantic populations and a distant Mediterranean population of barnacles is reported here for the first time. The Eastern Mediterranean (EM) cluster, that consists of Bodrum, Dubrovnik, Fažana, Larnaca and Rethymno, is well-defined in all our different analysis approaches. The Mid-Mediterranean (MM) cluster consists of Bastia, Birżebbuġa and Pantelleria. Bastia, at the northwestern coast of Corsica, belongs geographically to the Western Mediterranean basin. The other two MM locations are also close to the Western Mediterranean. Nevertheless, the MM cluster emerged as a sister group to the EM cluster in all our analysis approaches. While this is not surprising for Birżebbuġa and Pantelleria, the Bastia population is an integral part of this cluster. This finding is supported by a previous C. stellatus study, clustering Bastia, Genoa and nearby locations with MM populations and not with the Atlantic cluster (Pannacciulli et al., 1997). The only exception to this cluster analysis is presented in two of the 10 trees (Fig. 6) in which the population of Pantelleria is clustered with the WM.
Establishment of new barnacle populations, as well as other sessile animals, mainly occurs by current-assisted larval distribution (Johannesson, 1988). The pelagic stage of C. stellatus is about 22 days, allowing for extensive connectivity between populations (Pannacciulli, Manetti, & Maltagliati, 2009). The well-documented surface current, coming from the Atlantic Ocean and entering the Mediterranean through the Strait of Gibraltar can easily bring barnacle larvae to the shores of Málaga and Melilla. Previous studies have included these locations as part of the Atlantic region population, separated from the rest of the Mediterranean by the Almeria-Oran Front (Pannacciulli et al., 1997). But, according to our results, it appears that the Almeria–Oran Front is not impermeable to the propagules ofC. stellatus which is reflected in the resemblance of the Western Mediterranean populations of both sides of the front. Patarentello et al. (2007) reviewed over 20 population studies of 58 different marine species across the Atlantic-Mediterranean range, aiming to comprehend phylogeographical patterns, including potential barriers in the Mediterranean Sea. The patterns obtained from their data were very diversified, even between closely related species. The three major patterns were (i) full congruence between Atlantic and Mediterranean clades; (ii) distinct Atlantic and Mediterranean clades, where the Almeria–Oran front serves as the Atlantic–Mediterranean phylogeographical break; (iii) an Eastern Mediterranean clade that is distinct from the Western Mediterranean and Atlantic Ocean clade where the Sicily Strait and the Messinian Strait serve as a phylogeographical boundary (see also Villamor, Costantini, & Abbiti, 2014). Our results further expand the second and third models of Patarentello et al. (2007) and divide the populations to four clusters. In contrast to the previous above-mentioned analyses, that are based on a limited number of markers and small samples of specimens representing each population, our results that are based on a larger set of informative markers and a large population sample, look more reliable.
The Atlantic current, entering through the Strait of Gibraltar, extends east along the shores of Northern Africa (Hamad, Millot, & Taupier-Letage, 2006; Millot & Taupier-Letage, 2005; Poulain et al., 2013) (Fig. 8) elucidates the resemblance between Bizerte and the other WM cluster populations and the influence by the sister cluster of the EA populations. This Atlantic current system is likely to assist larval distribution entering the Mediterranean. It appears that the influence of this current does not spread further east and we can only find traces of the connection between the Bizerte and the Pantelleria populations (Fig. 6), which can also be explained by local gyres. The EM cluster is well defined and spans over a large part of the Mediterranean, from the eastern shores through the Aegean Sea and north into the Adriatic Sea. This part of the Mediterranean is influenced by the Asia Minor current as well as some cyclonic and anti-cyclonic gyres (Pinardi, Arneri, Crise, Ravaioli, & Zavatarelli, 2006) and we therefore witness increased connectivity between the different populations. The northern part of the Western Mediterranean is influenced by currents from the area of Sicily (Pinardi et al., 2006), and therefore it is not clear why the population resembles the MM populations and not a WM one like Bizerte. The mixing of water by the mesoscale gyres in the Tyrrhenian sea (Fig. 8) may contribute to the resemblance between the Bastia population and that of the other MM populations. Pantelleria is very close to the border between the well-defined branches of the EA/WM and the MM/EM, and indeed found to be genetically close to both. Although belonging to the WM cluster in most analysis, it is part of the WM cluster in two of the population trees, right next to its geographically neighboring population of Bizerte.
We may also speculate that the separation between the Western and Eastern Mediterranean populations of C. stellatus is a result of the geological history of the Mediterranean. One of the most conspicuous events that shaped the composition of flora and fauna of the Mediterranean is the Messinian Salinity Crisis (MSC) that started at the end of the Miocene, about 6 million years ago (Krijgsman, Hilgen, Raffi, Sierro, & Wilson, 1999) when the connection between the Mediterranean and the Atlantic was cut. The water balance of the Mediterranean was, and still is, deficient and is compensated by the Atlantic inflow through the Strait of Gibraltar. The Messinian Salinity Crisis ended in the Zanclean flood, occurred 5.33 million years ago when the Strait of Gibraltar opened and have refilled the Mediterranean Sea. During the MSC, the water level and salinity of the Mediterranean fluctuated and it was dried up and refilled repeatedly during the few million years of the Messinian stage, leaving behind lakes of different salinities (Hsu, Ryan, & Cita, 1973). Some of these lakes might be a refuge, inhabited by resistant organism like intertidal barnacles that naturally withstand fluctuating temperatures and salinities. We may propose that eastern and middle populations of C.stellatus are a relic of the MSC lakes while the WM populations resemble the Atlantic “invaders” arriving like many other organisms in the Zanclean flood.
In order to better understand how these populations were shaped, the selective pressures of the habitats, like salinity and temperature, must be considered. When compared to other species of barnacles (Bhatnagar & Crisp, 1965), C. stellatus was found to be better adapted to higher temperatures while less inhabiting lower salinity niches. Low salinity has been correlated with reduced number of eggs per brood inC. stellatus (Barnes and Barnes, 1965). These adaptive advantages can contribute to a bigger, stable and more polymorphic population in the higher salinity and temperature conditions of the EM. Indeed, the EM populations have more alleles per position and their expected heterozygosity is significantly higher than all other populations (Figs. 7a–b). The EA populations, on the other hand, have significantly smaller number of alleles per position, lower expected heterozygosity and lower percent of polymorphic positions (Figs. 7a–c), presumably reflecting the adaptive difficulties of C. stellatus in the lower salinity and temperature of the East Atlantic Ocean. The Atlantic current influence on the Mediterranean is also evident in the low salinity region stretching from the Strait of Gibraltar along the Northern Africa shores to Bizerte (Fig. 8). As expected from this, the WM populations do have less alleles per position and their expected heterozygosity is significantly lower than the EM populations. The environmental conditions may also explain the establishment of EM populations of C. stellatus in the western part of the Mediterranean. Mean surface salinity and temperature of the Northern Tyrrhenian and the Ligurian Seas are the highest in the Western Mediterranean Basin (UNEP/MAP 2012, p 23). These favorable conditions could allow for the establishment of a polymorphic population like the one in Bastia.
Patarentello et al. (2007) have concluded that genetic diversity does not necessarily decrease in a direction either from the Atlantic Ocean to the Mediterranean or even to the Adriatic Sea. The higher genetic variability of the EM population can be partially explained by the central-marginal hypothesis, also termed the Carson hypothesis (Carson, 1959; Eckert, Samis, & Lougheed, 2008; Sagarin & Gaines, 2002). The hypothesis claims that range margins exhibit less genetic diversity and greater inter‐population genetic differentiation compared to range cores. Since the Eastern Mediterranean shores are not a naturally occurring edge of the population this effect can only be seen in the Atlantic where C. stellatus does not spread further south to the Western Africa shore and even the Cape Verde Chthamalus was shown to be a different species (Tikochinski et al., 2020).
Yet another explanation may be the advantage of variability within the eastern basin populations in overcoming the higher temperature changes as well as other rapid condition changes and processes typical to this part of the Mediterranean, especially since the opening of the Suez Canal.
Our study presents a unique opportunity to study processes of population settlements in sessile animals, the influence of oceanographic conditions and processes including selection and genetic variation. In order to better understand and solidify some of our speculations, more studies of this range of geographic distribution, population size and genetic polymorphism are needed.