Amphibalanus amphitrite (Darwin, 1854)

Moeinadini, Asma, Sari, Alireza, Shahdadi, Adnan, Katouzian, Ahmad-Reza, Sarafrazi, Alimorad & Elahi, Elahe, 2023, Fig. 11. Left G 1 s in Tuerkayana latens, a New Species of Land Crab from French Polynesia, with a Discussion on the Phylogeny of the Genus (Crustacea: Decapoda: Brachyura: Gecarcinidae), Zoological Studies 62 (16), pp. 1-17 : 4-12

Specimens of Amphibalanus amphitrite

were collected from nine locations along the PG and the GO ( Table 1 , Fig. 1 ). In total, 94 individuals were collected from both artificial substrata (e.g., humanmade structures, piers, small vessel hulls and floating objects) and natural habitats (e.g., intertidal rocks, mollusk shells, crab carapaces and mangrove trunks). The specimens were kept in 96% or absolute ethanol immediately upon collection and transferred to the Molecular Systematics Laboratory at the University of Tehran for molecular analysis.

DNA extraction, amplification and sequencing

Total genomic DNA was extracted from muscles using the salt precipitation method ( Katouzian et al. 2016). A 576-bp fragment of the cytochrome c oxidase subunit I gene (COI) was amplified by the polymerase chain reaction using primer pairs LCO1490-JJ and HCO2198-JJ ( Astrin and Stüben 2008) as described in Chen et al. (2014). The PCR products were outsourced for sequencing to LGC Genomics GmbH (Berlin, Germany) and Macrogen Europe, Amsterdam using the same forward primers. Sequences were proofread via Chromas Lite (v. 2.1.1) (Technelysium Pty Ltd, Queensland, Australia). Sequences of all unique haplotypes/genotypes were submitted to GenBank (http://www.ncbi.nlm.nih.gov) and are available under accession numbers (OQ119797–OQ119890). COI sequences of A. amphitrite from previous study ( Chen et al. 2014) were also retrieved from GenBank and included in the analyses (accession numbers KC138445, KM211362–KM211497). Amphibalanus reticulatus , A. variegatus and Balanus glandula were used as outgroups (GenBank accession numbers JQ035518.1, JQ035522.1 and KU204282.1, respectively).

Sequence data analyses

Sequences were aligned using Clustal W ( Thompson et al. 1994; Villesen 2007) implemented in BioEdit 7.0.5 ( Hall 1999). The ML tree was obtained using raxmlGUI, v. 1.3 ( Silvestro and Michalak 2012) with 1000 bootstrap pseudoreplicates. The selected evolutionary model was GTR+G ( Rodriguez et al. 1990). Two maximum parsimony haplotype networks ( Templeton et al. 1992) were constructed with PopART ( Leigh and Bryant 2015), one for clade I at the global level ( Chen et al. 2014) taken from the tree, and one for the sequences of the present study.

We calculated the distribution of pairwise differences (i.e., mismatch distribution; Rogers and Harpending 1992) to trace population size change in DnaSP v.5.10 ( Librado and Rozas 2009).

To assess how mtDNA effective population size changed through time, we analyzed historical demography using coalescent based Bayesian Skyline Plot (BSP) in BEAST 2.4.7 ( Bouckaert et al. 2014). We selected GTR+G as the best model of nucleotide substitution and adopted a substitution rate of 3.1% per MY for COI (according to Tsang et al. 2008). We set a strict clock model as prior and ran three independent MCMC analyses with 60 million generations, sampling every 6,000 steps, to verify the consistency of the results. The initial 25% of the samples were discarded as burn-in. The convergence of all parameters was tested and BSP produced in Tracer 1.6 ( Rambaut et al. 2014).

Standard genetic indices were calculated to determine the genetic diversity within the species based on COI sequences. The haplotype diversity (h), nucleotide diversity (π), number of polymorphic sites (Np) and number of haplotypes (Nh) were calculated using DnaSP v.5 ( Librado and Rozas 2009) for each local population and for the complete dataset of each Gulf. Overall mean p ‐distance was analyzed with MEGA v.6 ( Tamura et al. 2013). We computed pairwise ΦST with 1,000 permutations in Arlequin v.3.5.2.2 ( Excoffier and Lischer 2010) to investigate population differentiation patterns among local populations (only populations with n ≥ 8 were included).

Species distribution modelling

To construct the distribution modelling of A. amphitrite , we included localities of specimens of clade I and III used in worldwide phylogenetic analysis compiled from field observations and available data based on Chan et al. (2014). The ocean climate layers were downloaded from Bio-ORACLE (Ocean Rasters for Analysis of Climate and Environment) data set ( Tyberghein et al. 2012). Variables were selected based on their ecological meaningfulness and contribution rate species in distribution model. To avoid the effect of high correlation among layers, we first examined all layers using OpenModeller 1.0.7 ( De Souza Muñoz et al. 2011) and then used the Pearson correlation method to obtain higher correlative layers (> 0.7). Layers with lower correlation (<0.7) were selected for further analyses: Temperature (Mean and Range), Phytoplankton Mean, Salinity (Mean and Range), Current Velocity Mean, Min., Max., Ltmax, dissolved oxygen range and Phosphate Mean. All the data were downloaded as raster format with a 5-arc-minute from Bio-ORACLE data set ( Tyberghein et al. 2012).

RESULTS

Five out of the nine sampling localities were located within the PG, and four were located in the GO ( Fig. 1 , Table 1 ). Here, we postulated two main populations, namely the PG and GO populations. 36 specimens from the PG and 58 specimens from the GO were studied. Each of the 94 COI sequences contained approximately 576 base pairs (bp), and none of these contained a stop codon. In total, 53 polymorphic sites were found. 136 additional sequences of previous studies on the species were obtained from GenBank and included into the dataset for tree and haplotype network construction.

Sequence data analyses

In the ML tree, three distinct and well separated clades were recovered ( Fig. 2 ). Sequences of the present study (green circles) are well distributed in the tree, with representatives in two clades. Most sequences were placed in clade I and only two specimens of the present study were significantly different from the others ( Fig. 2 ) namely one from the PG (Kandaloo) and one from the GO (Gwatr) ( Table 1 ). The phylogenetic tree showed 186 sequences, including 92 sequences of the present study placed in clade I. The haplotype network constructed for the clade I recovered 101 haplotypes, with 36 unique haplotypes for the PG and GO ( Fig. 3 ).

In the haplotype network ( Fig. 4 ) of the Iranian samples (clade I), some haplotypes were present in only one of the two gulfs, while other haplotypes were found in both gulfs. Except for some outlying haplotypes, the network is nearly star-shaped and most of the haplotypes remained close to the main haplotype with only one mutation step. However, the haplotype network ( Fig. 4 ) showed no clear patterns of isolation between specimens from the two gulfs.

The mismatch distributions showed a clear unimodal pattern in the populations of the PG and the GO. The distribution of pairwise haplotype differences was skewed to the left ( Fig. 5 ).

An exponential growth in effective population size, shown in Bayesian Skyline Plots ( Fig. 6 ), was consistent with the results of the mismatch distribution analysis. However, the BSP inferred slightly different timings of expansion from the mismatch analyses. The two populations shared a broadly similar timing of demographic growth, which began at 200 ka BP ( Fig. 6 ). The number of sequences available for the GO produced a plot showing a tendency to demographic expansion, truncated at 200 ka BP.

Diversity indices were calculated for material from the sampling localities with 10 or more sequences, as well as for each gulf in total. The analyses showed that the specimens from both gulfs had high genetic diversity, and that the diversity within the PG was higher than the diversity within the GO (i.e., haplotype diversity in the PG = 0.93; the GO = 0.87). The individual sampling localities of both gulfs also showed high genetic diversity ( Table 1 ).

The ΦST value was calculated only between populations with a sample size> 10 ( Table 2 ). Most pairwise ΦST values between populations were small, and only two pairwise comparisons were significant (p <0.05) ( Table 2 ).

Species distribution modelling

All models were run in ten replicates for the current time. The potential distribution models of A. amphitrite showed perfect Area Under Curve (AUC) test values, with an index of 0.89 ± 0.03 and 0.96 ± 0.03 for clade I and clade III, respectively. This showed significance for the binomial omission test, hence the maps were evaluated as very good (more than 0.850). The contribution performance of layers for each period is presented in table 3. Based on the results, temperature mean and range for clade I and temperature mean and current velocity min. for clade III made the largest contributions to current habitat suitability in the predictions modeling. Accordingly, the suitable habitats for A. amphitrite were the tropical and subtropical areas ( Fig. 7 ).

DISCUSSION

In the framework of biogeography, barnacles are among the most interesting invertebrates. During their pelagic larval stages, they are able to disperse under the influence of oceanic currents. By the end of the larval phase and after dramatic changes through metamorphosis, they become sessile which is expected to limit their active distribution, gene flow among populations, and cause decreased genetic homogeneity. T h e i r f o u l i n g b e h a v i o r, h o w e v e r, h a s g r e a t l y counteracted their natural dispersal ecology in recent decades ( Holm 2012). It seems that ships have played a great role in their dispersal throughout the open waters ( Yamaguchi et al. 2009). The fouling behavior increases gene flow among distant populations and therefore homogenizes their genetic structure; this in turn makes it difficult to pinpoint the origin of their establishment ( Ardura et al. 2016). One of the most common fouling barnacles is Amphibalanus amphitrite , an important species not only in marine ecology but also in the global economy ( Holm 2012).

Using the mitochondrial COI marker, Chen et al. (2014) studied global phylogeography of this species with material from 25 localities. They found three distinct morphologically similar clades (with a 4% divergence between clades) indicating historical isolation between populations or possibly a cryptic species complex. These include clade I, as the most common and globally distributed clade along tropics and warm temperate areas, clade II with only two sequences, one from Singapore and one from North Carolina ( USA) and clade III with more representatives, but restricted to the Indo-West Pacific tropics. It is commonly stated that the native range of A. amphitrite is the Indo-West Pacific ( Jones et al. 2000; Shahdadi et al. 2014). This is a huge water body with high diversity that is occupied by all three clades. However, centuries of hitchhiking with human-mediated vectors or corridors has made it difficult to determine the autochthonous population distribution throughout the Indo-West Pacific region (Carlton 2011). The study of Chen et al. (2014) also included two localities in the northwest of the Indian Ocean, namely Mumbai (East of India) and Saudi Arabia (the Red Sea), which were dominated by representatives of clade I, with only one sequence of clade III in Mumbai.

In the previous studies, no specimens from the Persian Gulf (PG), a major commercial and oil transportation destination, were investigated. In the present study, in addition to the sequences of the specimens collected from the PG and the GO, the sequences reported in Chen et al. (2014) were included in the phylogenetic analyses. The distribution pattern of the PG and the GO samples was similar to that previously reported for samples from the east of India ( Mumbai in Chen et al. 2014); most specimens were placed in clade I and only one sequence from each gulf was placed in clade III ( Fig. 2 ). Clade I specimens included barnacles from Japan to Malaysia and also from Australia, India, Saudi Arabia, Hawaii, California and North Carolina ( Chen et al. 2014). Within clade I, the PG and the GO samples clustered with sequences from Australia, West India, Taiwan, Hong Kong, Vietnam, Salton Sea ( USA) and Singapore.

Clade III was only present in tropical Indo-West Pacific waters ( Chen et al. 2014). In the phylogenetic tree ( Fig. 2 ), only two specimens from current study showed great affinity with clade III together with sequences from Australia, Taiwan, Hong Kong and Singapore. This clade had limited distribution compared to clade I. In the case of a complex species, it is expected to have one species or subspecies with a cosmopolitan distribution and other specimens restricted to specific areas ( Chan et al. 2007; Zhan et al. 2010; Keshavmurthy et al. 2013). We demonstrated that this clade is mostly distributed and mapped on the shipping routes. The results of the current study clearly confirm that anthropological, mainly shipping activities, are the most likely explanations for the present global distribution of A. amphitrite . This information is expected to help shipping policy makers in the designing of effective measures for protection of ships as well as in the maintenance of the biodiversity of barnacles ( Carlton et al. 2011; IMO 2021).

The spread of A. amphitrite may appear unidirectional toward the PG depending upon vectors, local hydrographic conditions and exploration history of larval stages. The vectors are important in transferring this species especially to the regions with high shipping traffic such as the PG and the GO. The PG produces around 46% of the world’s oil consumption, and more than 90% of this oil is transported annually in thousands of oil tankers ( Haapkylai et al. 2007; Al-Yamani et al. 2015).

The dispersal pattern of barnacle larvae and consequently the distribution range of the adults, are expected to reflect the interaction of larvae with oceanographic currents ( Keith et al. 2011; Tsang et al. 2012). The spatial structure of a species may mirror the major oceanographic systems and environmental conditions. Additionally, the Indian Ocean tsunami off the west coast of northern Sumatra, Indonesia in 2004 ( Lay et al. 2005; Okal et al. 2006) and, the Pangandaran tsunami off the west and central coast of Java, an island in the Indonesian archipelago in 2006 ( Fritz et al. 2007), affected major parts of the Indian Ocean. This resulted in the transport of vast bulks of water within only a few days to different parts of the Indian Ocean including the GO and the PG.

The currents of the PG and the GO are also anticipated to play important roles in determining the present dispersal pattern of A. amphitrite ( Ghanbarifardi et al. 2018; Sepahvand et al. 2021). For example, the anticlockwise currents that circulate from the GO to the PG ( Yao 2008; Thoppil and Hogan 2010) most likely facilitate the transfer of the species’ larvae (larval duration 7–17 days in 20–26°C) in the planktonic stages ( Costlow and Bookhout 1958). These passive transports will spread individuals of A. Amphitrite to the PG through different means, namely, natural larval distribution, transport and release of larvae by ballast water or carrying ovigerous barnacles on ship hull to new regions.

Temporal patterns of introduction of A. amphitrite to the PG and the GO should take into account dates of first records. The increased diversity of barnacle invasions in the last half of the twentieth century is in close synchronization with general global observations of increasing invasions of marine invertebrates, fish, and algae after World War II associated with vastly expanded global trade facilitated by more, larger, and faster ships. For example, A. amphitrite together with A. improvisus were the first barnacles introduced to Americas for the first 100 years of invasion history at 1853–1955 ( Carlton et al. 2011). In California, the first record of A. amphitrite is in 1914 ( Henry and McLaughlin 1975).

The first record of A. amphitrite from the PG was given by Nilsson-Cantell (1938) as Balanus amphitrite hawaiiensis from an unknown locality and then by Stubbings (1961) as B. amphitrite var. communis , and B. amphitrite var. hawaiiensis , from Kuwait; by Utinomi (1969) as B. amphitrite from Hormoz Island; and by Jones (1986) as B. amphitrite var. communis from Kuwait. A recent checklist of Iran barnacles recorded A. amphitrite from all over the coasts of Iran ( Shahdadi et al. 2014). Introduction of this species to the PG and the GO fell within a well-known global pulse of invasions related to an earlier surge of shipping, particularly in special economic zones. This widespread distribution is more related to clade I. Comparatively, clade III from the Indo-West Pacific has not expanded geographically. This distribution pattern reveals the major role of human-mediated vectors and corridors in introduction and dispersion of barnacles.

Present results revealed high levels of genetic diversity in the mitochondrial marker COI in the PG (h = 0.9) and the GO populations (h = 0.8). These values are also notable when compared to the population of several other invertebrate species in the PG and the GO, such as the fiddler crab Uca sindensis (h = 0.63; Shih et al. 2015), Jinga shrimp Metapenaeus affinis (h = 0.0– 0.33; Tamadoni-Jahromi et al. 2016) and pearl oyster Pinctada radiata (h = 0.0–0.47; Al-Saadi 2013).

Some other barnacle species from other parts of the world also demonstrated similar levels of genetic diversity in COI such as Amphibalanus improvisus (h = 0.75–0.96; Wrange et al. 2016), Chthamalus proteus (h = 0.9–1; Zardus and Hadfield 2005) and Tetraclita serrata (h = 0.9; Reynold et al. 2014).

The PG is a geologically young, semi-enclosed basin with a harsh environment, including high temperature and salinity. These extreme indices are believed to be largely responsible for the lower biodiversity of the PG ( Naderloo 2017). The conditions of the PG may allow a recent selective sweep for fixation of particular haplotype. In contrast, according to Shahdadi et al. (2014), the number of barnacle species in the PG is 33, which compares to 26 species in the GO. The A. amphitrite populations from two gulfs share several haplotypes including the most common one. The genetic diversity in the PG population is higher than the diversity of the GO and many other previously studied populations from around the world (see Table 1 in Chen et al. 2014).

Owing to the young age of the PG, the A. amphitrite population seems to have been introduced to this region naturally. This establishment was before species discovery by Darwin (1854). Its transport into the gulfs by ships occurred later. Therefore, according to common hypothesis, this population is expected to present low genetic diversity compared to native population. Our data revealed high genetic diversity among the PG barnacles. This is indicative of multiple episodes of introduction of barnacles to the PG from different sources (such as traditional Indian and Chinese shipping, tsunamis, tropical cyclones, and earthquake induced waves).

The high genetic diversity within the PG population and presence of common haplotypes among the PG and other populations of the world are therefore likely to be partially resulted from the anthropogenic introduction of haplotypes from around the world. The same reason could justify the high genetic diversity in some previously studied areas like Singapore and Hong Kong (see Chen et al. 2014). This could be well explained by the fact that ships from all over the world travel to these waters, stay in harbors for some time and subsequently transfer barnacles with different haplotypes from multiple sources to this area, providing their larvae with enough time for settlement and permanent sessile life in the region.

In this case, A. amphitrite has experienced multiple modes of dispersal including transportation of barnacle larvae by currents and ballast water and transportation of adults on ships hulls as fouling organisms to other parts of the globe. This could explain the presence of some of these newcomer haplotypes that are very distinct from the local haplotypes. However, intensive human-mediated gene flow has probably counteracted these processes. The result of all has been colonization of new areas at different scales and the presence of the dominant species on most coasts in the PG and the GO ( Shahdadi et al. 2014; Shabani et al. 2019; Al-Khayat et al. 2021). Transport has been mediated via most natural and anthropogenic substrates including floating marine debris of various types and surface texture. For example, plastics have been common and pervasive anthropogenic debris in marine environments ( Rech et al. 2018) and as floating objects they provide opportunities to alter the abundance, distribution and invasion potential of this species for effective colonization ( Goldstein et al. 2014). These transport opportunities may have been the key elements for successful distribution of this species. The knowledge about different types of dispersal dynamics of non-native species is crucial to our understanding of both evolutionary aspects of colonization as well as future management of biological introductions and invasions. Human-mediated dispersal has increased transmission of haplotypes from all over the world to these regions. This has caused increased genetic diversity within populations. On the other hand, with increasing gene flow, genetic differentiation has decreased as evidenced by genetic divergence (ΦST) between sampling sites that generally do not exceed 3% ( Table 2 ). The values suggest high gene flow and absence of or weak genetic subdivision between the populations of the biogeographical areas (see Guo and Wares 2017).

The smaller values of ΦST observed in the present study may derive from higher gene flow between individuals of A. amphitrite from different localities. The real distribution and connectivity of intertidal animals in the region and closely related areas (i.e., West Indian Ocean) are evidently determined by oceanographic regimes, environmental conditions and historical events ( Tsang et al. 2012; Afkhami et al. 2016; Rahimi et al. 2016; Ghanbarifardi et al. 2018).

The unimodal right skewed mismatch distribution curves ( Fig. 5 ) confirm the recent population expansion for individuals collected both from the PG and the GO. The Bayesian Skyline Plot analysis (BSP) displayed the demographic expansion curve for the populations of the PG and the GO areas with truncation possibly due to the smaller sample size ( Fig. 6 ). Despite more truncation for the GO, the curve shape was consistent with that of the PG. The results uncovered that the demographic expansion of A. amphitrite in the PG and the GO areas started approximately before 200 ka BP ( Fig 6 ). This dating could not be in accordance with the history of the PG, which appeared about 18 ka years ago, and the fact that just about 8–10 ka BP, the northeastern margin of the PG approached its present position in several localities ( Lambeck 1996; Marko et al. 2010). Thus, the expansion time of A. amphitrite in the PG and the GO had likely been at the early Holocene, concurrent with the re-flooding of the PG. This historical data seems to be related to gene flow in the Indian Ocean prior to the appearance of the PG. Consequently, it is not possible to precisely define the beginning of population expansion for this area. However, it is plausible that human factors played a role in introducing haplotypes to this region in recent decade.

Distribution modelling maps showed the presence of potential localities for geographical distributions of two clades based on natural environment factors. From this perspective, both clades are common in the China Sea, Northern Australia especially the Gulf of Carpentaria, PG, GO, the Red Sea, West African coastal areas, the east coast of the United States and the western coasts of Mexico. These are potential distribution locations for both clades, but clade III in the Indo-Pacific Ocean presents a noticeably broader distribution. With ever growing human intervention ( Seebens et al. 2013), we can expect changes in distribution of A. amphitrite clades in the future.

The shape of the population genetic structure of marine invertebrate species, in addition to historical demography and selection regimes, is affected by environmental factors ( Bohonak 1999; O’Riordan et al. 2004). According to previous records, A. amphitrite is widely distributed globally around tropical and subtropical coasts ( Henry and McLaughlin 1975; Chen et al. 2014). In the present study, distribution modelling showed the importance of environmental factors such as temperature, phytoplankton availability and salinity in the distribution of this species. Distribution of the fouling species is related to fouling type. The extent of fouling on ships’ hulls depended on many factors, e.g., water salinity, light, temperature, pollution, geographical location and nutrient availability ( Pettengill et al. 2007; Hulme 2009). Based on studies on invasions in marine environments ( Gallardo et al. 2015), the distribution of invasive species is influenced by nutrients (40%), temperature (26%), human footprint (21%) and other factors (13%) that support a relevant role of the match between physicochemical characteristics of the donor and receiving waters for species introductions ( Seebens et al. 2013). It seems that temperature plays the most important role in species distribution, specifically in tropical or subtropical waters. The polar zones are subjected to the most severe fouling attack, particularly in more shallow, coastal waters where there is greater abundance of light, heat and nutrients, resulting in prolific reproduction of the fouling species ( Ubagan et al. 2021). One of the important factors in future distribution of A. amphitrite is climatic warming scenarios ( Carlton 2000). The southward spread of A. amphitrite from Brazil to Argentina is a remarkable example ( Carlton et al. 2011), and therefore, distribution maps of modelling can help to understand the role of potential climatic changes on expansion ( Howard 1997; Poloczanska et al. 2008; Southward 1991).

The species displays several life history traits including broad environmental tolerance such as with temperature ( Qiu and Qian 1999; Piazza et al. 2016). In their study, Khosravi et al. (2019) suggested that during global warming, A. amphitrite has maximum biofouling coverage in the PG and overtakes more rivals. Furthermore, this species can survive in the high salinity of the PG ( Simpson and Hurlbert 1998) better than the closely related mesohaline barnacle, A. improvises . The latter species is often restricted to low salinity environments and does not show a structured population genetic pattern globally ( Wrange et al. 2016). However, A. amphitrite has displayed three clades with potentially different physiological properties which display distribution restrictions (Chan et al. 2014). This tolerance ability has granted A. amphitrite a competitive advantage over rivals. Remarkably, Bishop (1947) reported that three species of live barnacles, including A. amphitrite , arrived in Liverpool, England, after a 30- days voyage from Australasia, via the Panama Canal.

Figures

Fig. 1. Map of the sampling localities.

Fig. 2. Phylogenetic tree constructed with Maximum Likelihood (ML) based on the COI gene for selected sequences of A. amphitrite, focusing on Iranian populations (green circles). Amphibalanus reticulatus (JQ035518.1), A. variegatus (JQ035522.1) and Balanus glandula (KU204282.1) were used as outgroups. Numbers show the bootstrap values after 1000 pseudo-replicates. Clades I, II and III are corresponding to those in Chen et al. (2014).

Fig. 3. Maximum-parsimony mitochondrial DNA haplotype networks for material from the PG and the GO, and the retrieved sequences from the GenBank (only sequences of the clade I in Chen et al. (2014) phylogenetic tree are included).

Fig. 4. Maximum-parsimony of mitochondrial DNA haplotype networks for the PG and the GO populations constructed in PapArt. Hatch marks represent mutations; numbers in the figure legend refer to sampling sites in table 1 and figure 1.

Fig. 5. Frequency distribution of the number of pairwise nucleotide differences (mismatch) between COI haplotypes in the two populations of Amphibalanus amphitrite. The solid line shows the theoretical distribution under the model of demographic expansion.

Fig. 6. Bayesian skyline plots of effective population size through time in Amphibalanus amphitrite from two biogeographical areas (A: PG and B: GO), based on the 576 bp sequences of COI and a nucleotide substitution rate of 3.1%/MY. The bold black curve is the median of the parameter NeT, which is proportional to the effective population size; the blue lines delimit the 95% highest posterior density. For comparison, all x-axes have the same scale. The plots are truncated to the median estimate of each area’s TMRCA.

Fig. 7. MAXENT reconstruction for clade I (A) and clade III (B) of Amphibalanus amphitrite in the world representing current distribution models.