Model selection

The best substitution model according to the AIC was TVM + G + I, while the best model according to AICc and BIC was TPM3uf + G + I (see jModeltest manual). These models were implemented in GARLI analyses, but they are not available in RaxML and MrBayes. In addition, the weights for these models were not high, particularly for the AIC and AICc (see Table 1) indicative of model selection uncertainty. The 99% cumulative weight interval included the most complex model evaluated (GTR + G + I). We therefore implemented the GTR + G + I, as well as the simpler GTR + G in RaxML, MrBayes, and GARLI, because of potential problems with estimating G + I simultaneously [see RaxML manual and pages 113–14 in 43].

Phylogenetic results

Our phylogenetic analyses included all of the localities in Figure 2 as well as samples from all the Ligia spp. with sequences from both genes available in GenBank. Ligidium hypnorum, also in the family Ligiidae, was used to root our trees. Our results recovered a relatively well supported monophyletic group (Clade I; Fig. 3) that contained all of the Ligia sp. samples collected between central California and central Mexico, L. exotica samples from Hawaii and South Carolina, L. perkinsi (Hawaii), L. hawaiensis (Hawaii), and L. vitiensis (Madagascar), to the exclusion of samples of L. oceanica (Germany), L. pallasi (Canada and USA), and L. italica (Italy). Within Clade I, L. perkinsi, L. hawaiensis, and L. vitiensis appear as a sister clade (Clade III) to all the remaining samples (Clade II), although node support values for the reciprocal monophylies of Clades II and III were highly variable. Within Clade II, five monophyletic groups were recovered: central California clade (Clade A; grey in Figs. 2–3); Baja Pacific-Southern California clade (Clade BCDE; green, orange, magenta, and turquoise in Figs. 2–3); Gulf clade (Clade NS; red and blue in Figs. 1–3); Careyes clade (Clade F; brown in Figs. 2–3); and L. exotica. The relationships among the five clades within Clade II (i.e., A, BCDE, NS, F, and L. exotica) were unresolved. Relationships and phylogeographic patterns within each of these clades (except L. exotica) are described below.

The central California clade (Clade A; grey; Figs. 2–3) was comprised of mainland samples from the vicinity of the San Francisco Bay area and samples collected in the Northern Channel Islands. These island lineages are mainly located in the western portion of the Northern Channel Islands (west Santa Rosa and San Miguel), with the exception of Orizaba (A4) in Santa Cruz Island. Very shallow divergences were observed within this clade (maximum Kimura-2-parameter distances were 0.37% and 1.55%; for 16S rDNA and COI, respectively; Table S2).

The Baja Pacific–Southern California clade (i.e., BCDE) contained all the other samples collected from California and all the samples from the Baja Peninsula Pacific side. This clade was first divided into two groups: a group that contained samples from Santa Catalina Island and mainland southern California localities from San Diego to Refugio (“Southern California”; Clade B; green; Figs. 2–3); and a group that contained samples from the Baja Pacific, northern Channel Islands, and San Diego (i.e., Clade CDE). The CDE clade in turn was divided into two groups (i.e., E and CD). The first one (Clade E; “Baja Pacific South”; turquoise; Figs. 2–3) contained all of the Baja Peninsula Pacific localities south of Guerrero Negro Lagoon. The second one (i.e., CD; orange + magenta) contained all of the localities on the Baja Peninsula north of Guerrero Negro, a sample from San Diego, and several localities from the Northern Channel Islands. Clade CD was divided into a group that included only localities from the northern Pacific Baja Peninsula (Clade D; Baja Pacific North; magenta; Figs. 2–3), and a group that contained mainland localities from Ensenada (Mexico) and San Diego, and Northern Channel Islands localities (Clade C; “California”; orange; Figs. 2–3). These island localities are distributed in the eastern portion of the Northern Channel Islands (Anacapa, Santa Cruz, and eastern Santa Rosa).

Maximum sequence divergences within each of the B, C, D and E clades reached 2.22, 0, 0, 2.41%; respectively, for the 16S rDNA gene (Table S2), whereas they reached 8.60, 2.10, 2.08, 8.77%; respectively for the COI gene (Table S2, S3, S4, S5). Divergences among clades B, C, D and, E ranged between 1.02–7.5% for 16S rDNA and between 7.28–19.6% for COI (Table S2).

The Gulf clade (Clade NS; red and blue; Figs. 2–4); grouped all the localities that were sampled in the Gulf of California and all mainland samples collected south of the Gulf, with the exception of Vallarta and Careyes (brown in Figs. 2–3). The Gulf clade was divided into two well-supported main lineages: Gulf North clade (clade N; red; Figs. 2–4) and Gulf South clade (clade S; blue; Figs. 2–4). Divergences between the Gulf North and Gulf South clades ranged between 5.32–11.09% for 16S rDNA and between 15.16–26.47% for COI (Table S2). COI divergences among selected localities/lineages within the Gulf North and Gulf South clades are shown in Tables S6 and S7; respectively. The Careyes clade (Clade F; brown) included only the samples from Vallarta and Careyes.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Maximum Likelihood tree of Ligia Gulf samples (expansion of the Gulf clade in Figure 3).

Taxon IDs, clade colors and shapes correspond to localities in Figure 2. Numbers by nodes indicate the corresponding range of node support values obtained for each method: Top-Bayesian Posterior Probabilities; Middle-GARLI bootstrap support; and Bottom-RaxML bootstrap support. * denotes nodes that received 100% support for all methods. Nodes receiving less than 50% support for all methods were collapsed. Nodes with no corresponding support values were of little relevance or had low support values.

https://doi.org/10.1371/journal.pone.0011633.g004

The Gulf North clade (Clade N; red) included: (1) Baja samples from San Bruno and Santa Rosalia (in the mid-peninsula), and all Baja localities sampled north of Santa Rosalia; (2) all Gulf mainland localities sampled from Puerto Libertad (N16) to the north; and (3) all samples collected from the central Gulf islands (Tiburón, Angel de la Guarda, San Esteban, San Pedro Martir, El Cholludo, Isla Datil; N3 and N7–N8). Within the Gulf North clade, a well-supported group (red circles; N10–N14; Figs. 2 and 4) was formed by localities from the upper Gulf between Puerto Peñasco (mainland), and San Luis Gonzaga (Baja), which was sister to a well-supported group formed by Puerto Lobos and Puerto Libertad (red circles; N15–N16). The sister lineage to this “red circles” clade (N10–N16) appears to be Angel de la Guardia Island (N6), but support for this relationship was variable. The island localities of Tiburon, El Cholludo and Isla Datil (N7) formed a well-supported group, whose closest relatives appear to be San Esteban Island (N8) and San Rafael (a Baja Peninsula locality; N9). San Francisquito, San Pedro Martir Island, Santa Rosalia and San Bruno formed a well-supported group (red squares N1–N4). The relationship of Bahía de los Angeles (N5) with the other lineages was not resolved.

The Gulf South clade (Clade S; blue) included: (1) San Lucas and Santa Rosalia (S7) and all populations south of San Bruno in the Baja Gulf; and all mainland populations collected from Kino (S9) to central Mexico (with the exception of Vallarta and Careyes). Baja Peninsula localities north of the Cape region (i.e., North of La Paz) were grouped into two well supported main clades: one (blue hexagons) included the localities sampled from Santa Rosalia and San Lucas to Bahía Concepción (S6), excluding San Bruno; the other (blue stars) included the localities from San Brunito to El Cajete. Baja Cape region lineages were grouped into three separate lineages: one that included all the localities from La Paz and Isla Espiritu Santo (blue diamonds); one that included all Baja Cape region localities south of La Paz with the exception of Cabo Pulmo (blue triangles); and one that included Cabo Pulmo (blue “X”; S26). These lineages formed a monophyletic group (i.e., “Baja Cape-Southern Mainland” clade; Fig. 4) with a clade comprised of all the mainland localities in the southern Gulf and those south of the Gulf (blue circles), with the exception of Vallarta and Careyes. The sister to the “Baja Cape-Southern Mainland” appears to be a clade (blue squares) formed by the central Gulf mainland localities of Kino, Guaymas, and San Carlos (S9–S10) and the central Baja locality of San Nicolas (S27).

Taxonomic uncertainty and Central California Clade

Our inferred phylogenetic relationships among Ligia isopod samples from central California to central Mexico show a rather complex and previously unknown evolutionary history. The high degree of sequence divergence among several lineages suggests that multiple Ligia species exist. Recognized species of Ligia show COI divergences starting at 14% (L. hawaiensis vs. L. perkinsi; Table S2). Many of the lineages found in our study area are at least 14% divergent from their closest relative (Tables S2 and S6–S7), although smaller divergences may also represent different species. Accordingly, the taxonomy and systematics of Ligia in this region need to be revised. Traditionally, with the exception of L. pallasi in central California, one species has been recognized in this region, Ligia occidentalis; although some authors also indicate the presence of L. baudiniana (a species from the Atlantic) and L. exotica (a cosmopolitan species, but see discussion below) [15], [16]. The type locality of Ligia occidentalis Dana, 1853 is around the San Francisco area we sampled; thus, our samples from this region likely correspond to this species. These samples clustered into the Central California Clade (Clade A; grey), with several samples from the Northern Channel Islands. The small divergences observed among the localities within this clade (Table S2) suggest a recent exchange between the mainland and the islands. However, this clade is highly divergent from the other main lineages (Table S2), suggesting they correspond to different species. The regions between Point Conception and San Francisco and north of San Francisco need to be explored to determine the distributional limits of clades A, B, and C.

The four main lineages found in our study area (i.e., clades A, BCDE, Gulf, and Careyes) are closely related to L. exotica, L. hawaiensis, L. perkinsi, and L. vitiensis. In contrast, L. pallasi, a coastal species whose distribution overlaps with our study area (distributed between Santa Cruz, California, and Alaska), and for which we included samples from Canada and Washington, USA, is very distant from the Ligia examined in this study. Ligia exotica has been considered a single exotic species, commonly found in harbors and ports around the world, and thus regarded as introduced by ship traffic in many areas [reviewed in 15]. However, molecular analyses of samples from different parts of the world also show deep divergences among lineages, suggesting L. exotica is a complex of cryptic species (Hurtado et al., unpublished). Similarly, L. hawaiensis and L. perkinsi, both from the Hawaiian archipelago, show deep interpopulation genetic divergences, but the phylogenetic relationships among populations of the two members have not been well resolved [11]. To resolve the sister relationships of Ligia lineages found in our study area, a comprehensive examination that includes representatives of potential sister lineages and additional genes is necessary.

Baja Pacific – Southern California (BCDE) Clade

Another deep lineage includes the clades B, C, D, and E. In the Southern California clade (Clade B; green), mainland localities from Los Angeles area (B5, B8, B9) to Refugio Beach (B7) formed a well-supported group. The localities sampled around the Los Angeles region (i.e., Long Beach, Malaga and Cabrillo) have identical 16S sequences suggesting recent exchange, which may be a consequence of the relatively young age of the current rocky headlands in southern California [1]. This group plus Refugio split from the rest of the members in the Southern California clade (B), which includes samples from San Diego and Santa Catalina Island, suggesting a vicariant event around Los Angeles region (represented by 6.89–8.04% COI divergence; Table S3). Remarkably, interspecific phylogeographic breaks have also been identified in the Los Angeles region for other intertidal and coastal animal taxa [44], implying they may share a common vicariant history in this region. The small divergences (0.9–2.42% COI; Table S3) observed between Santa Catalina Island localities (B1–B3) and the mainland (i.e., San Diego B6) may indicate a recent exchange, consistent with the young age of the island (∼1 Mya), which is suggested to have separated from southern California [45].

Within the California-Baja Pacific North-Baja Pacific South clade (i.e., CDE clade), a mid-peninsular phylogeographic break occurs at the Guerrero Negro Lagoon area (i.e., California + Baja Pacific North vs. Baja Pacific South). Thus, presence of this Lagoon may be associated with this phylogeographic split. Remarkably, a mid-peninsular phylogeographic break is also observed in Ligia lineages along the Baja peninsula inside the Gulf (i.e., Gulf North vs. Gulf South clades). A seaway may be the cause of the breaks on both sides of Baja (but see Gulf Clade Discussion).

Divergence of the Baja Pacific North Clade (D; magenta) from the California Clade “C” (orange) may be associated with one of the multiple marine incursions that are suggested to have isolated the peninsula from mainland between 5–14 Ma [reviewed in 46]. Within the California Clade C (orange) the small COI divergences observed (maximum 2.1%; Tables S2 and S4) among the northern Channel Island localities and between these and their closest relatives on the mainland (i.e., C1 and C2), suggest recent exchange among these localities (similar to the Channel island populations of the Central California clade A; grey). These islands appear to have formed ∼3 Ma [N. Pinter pers. comm. in 47]. However, Anacapa and San Miguel islands were completely under water during at least one interglacial episode [45], so Ligia populations present before would have gone extinct. Santa Rosa and Santa Cruz islands have been continuously above sea level for at least the last 500,000 years [45]; but it is possible that they were inundated before. The four present-day Northern Channel Islands were all part of a large contiguous land mass ∼17,000 years ago, which is believed to have been connected to the mainland and facilitated dispersal of terrestrial animals to this island [45]. Based on the distribution of non-vagile reptiles, it has been also suggested that the Northern Channel Islands were once connected to mainland, near present-day Mexico, and were carried farther north along fault systems [45]. The grouping of these Northern Channel Islands lineages with samples from San Diego and Ensenada suggests an origin of the island lineages around the Mexico-US border. Interestingly, the distribution of the two main lineages found in the Northern Channel Islands is segregated geographically: California clade C lineages (orange) in the east; and Central California clade lineages (grey) in the west, which may be related to the past history of colonization and fragmentation of the islands.

Experimental reciprocal crosses provide further evidence for multiple species of Ligia in California [16]. Crosses between individuals from Los Angeles and Carpenteria (found between B7 and B8; Fig. 2) produced normal offspring. However, no offspring were observed in crosses of Carpenteria and San Francisco, which were found to be highly divergent in our study; as well as between Carpenteria and the Channel Islands of San Nicolas and San Clemente, which were not sampled in our study.

Gulf Clade

The Gulf clade is characterized by extremely high genetic divergences among and within several lineages (north vs. south COI divergences: 15.16–26.47%; within north: up to 25.3%; and within south: up to 21.55%; Supporting Tables S2 and S6–S7), suggestive of a long history in the region. Due to the high allopatric divergences among the lineages within this clade, its apparently long history in the region, and a broad geographic distribution in the Gulf and adjacent areas, Ligia has the potential to have retained in its phylogeographic patterns, signatures of past vicariant events in this region. However, interpretation of this pattern is limited by the incomplete and controversial state of knowledge on the complex geological history of the Gulf of California and Baja California Peninsula; for which alternative and strikingly different hypotheses have been proposed [e.g. 21],[22]–[27].

The mid-peninsular breaks observed for Ligia at the Pacific (i.e., Guerrero Negro area) and Gulf (i.e., Santa Rosalia area) sides of Baja appear to be consistent with the presence of a seaway, because such a barrier would likely cause north-south breaks at both sides of the peninsula for a coastal organism. However, the mid-peninsular break in the Pacific is shallower than the one observed in the Gulf, although substitution rate heterogeneity, which is observed in our dataset (P<2.05×10⁻¹⁷; d.f. = 130), could account for the different divergences. Although the break around Santa Rosalia is consistent with the proposed location for a seaway opening, the break in the Pacific is ∼170 Km north of the suggested location (black circle; Fig. 2) for a seaway opening in the Pacific [48]. Therefore, it is unclear whether or not the Pacific and Gulf mid-peninsular breaks are associated with the same event.

Existence of a mid-peninsular seaway has been a contentious issue in phylogeographic and geological studies. A ∼1 Ma mid-peninsular seaway was originally proposed to explain a mid-peninsular divergence of Uta lizards [28], but the same research group indicates in subsequent papers that this divergence may represent Late Miocene–Early Pliocene times [24], [49], [50]. Mid-peninsular phylogeographic breaks have been reported in multiple taxa [e.g., 22]; but a broad range of divergences is observed, with some showing deep [46], [49], [51], while others shallower phylogeographic breaks [22]. Riginos [52] suggests that a Plio-Pleistocene mid-peninsular seaway is the simplest explanation for a concordant genetic division within both terrestrial and marine vertebrates. However, a comparative phylogeographic analysis using data from multiple studies concludes that two mid-peninsular diversification events occurred [53]. In addition, Grismer [54] indicates that mid-peninsular divergences may be associated with abrupt habitat and climate changes in central Baja, rather than a seaway. No geological evidence for a mid-peninsular seaway ∼1 Ma exists [24], but geological evidence (albeit controversial) suggests the existence of a more ancient seaway. Based on characteristics and distributions of fossil assemblages and marine deposits, Helenes and Carreño [26] suggest a Miocene seaway connected the Pacific with a northern proto-Gulf basin (discussed below) through the central part of Baja. However, Oskin and Stock [55] consider that a seaway does not necessarily explain the distribution of Miocene marine deposits in Baja. Nevertheless, paleomagnetic data indicate the presence of a seaway in Santa Rosalia ∼7 Ma [56].

Regardless of whether or not a seaway was the cause for the Gulf mid-peninsular break in Ligia, it does not appear to be a ∼1 Ma event. The high divergence between Gulf North and Gulf South clades (COI Mean  = 21.68%±1.9 SD) suggests a divergence time older than 1 Ma, probably in the Miocene, unless very elevated substitution rates occur in Ligia, which is unlikely. The mutation rate of COI in other marine isopods is suggested to be 2.5%/My [57]; thus, for this divergence to represent 1 My, it implies Ligia has a substitution rate ∼9 times higher than that reported for other marine isopods.

Other aspects of the phylogeographic patterns of Ligia are incongruent with the hypothesis of Riddle [i.e., Fig. 1; 22]. Under Riddle's hypothesis, reciprocal monophylies of the mainland vs. the peninsula (or at least the monophyly of one of them) are expected. However, within the Gulf, Ligia lineages of southern Baja are more closely related to mainland lineages (south of ∼29°N latitude) than to northern Baja and northern mainland lineages. Furthermore, Ligia Cape region lineages are more closely related to southern mainland lineages (south of ∼25°N latitude) than to any other Baja lineages. This is in striking contrast to the suggestion that the Baja Cape region was the first part of the peninsula to separate from mainland [22], [55]. Dispersal of Ligia across long stretches of ocean between southern Baja and mainland may explain these discrepancies. We believe this is unlikely because if Ligia from the Gulf of California had such dispersal abilities (i.e., peninsula–mainland distances are >100 Km in the central Gulf and >190 Km in the mouth), we would not expect to see the exceptional degree of allopatric differentiation exhibited. Furthermore, mixing of distant localities by dispersal across long stretches of ocean would likely have produced a very random phylogeographic pattern, which, in general, does not appear to be the case. Nevertheless, such dispersal cannot be completely ruled out. However, the phylogeographic patterns of Ligia in the Gulf clade appear to be congruent with elements of alternative geological hypotheses that consider an older history for the Gulf, as well as with phylo- and biogeographic patterns of other taxa.

The existence of an isolated Late-Miocene proto-Gulf basin that included the northern portion of today's Gulf and an extensive area to the north is well-accepted and was proposed since the 1970's [25], [58]–[61]. Recent evidence suggests that this northern proto-Gulf is at least 11.61 Ma old [60]. As mentioned above, a connection between the proto-Gulf and the Pacific appears to have existed in the central part of the peninsula [25], [26], [48]. How the Gulf evolution proceeded from the northern proto-Gulf stage to the present-day Gulf is unclear [25]. Nevertheless, several studies suggest that the southern part of Baja separated from mainland more recently than northern Baja, and that the Cape region was the last part of Baja to separate from mainland ∼4–6 Ma [25], [26], [61]. Such sequence of events is consistent with the phylogeographic patterns of Ligia in the Gulf South clade. Ledesma-Vásquez [62] proposes that a southern basin formed, younger (3.5–5.5 Ma) than the northern proto-Gulf, and that the older northern proto-Gulf joined this southern basin to form the present-day Gulf; although the exact sequence of events is unclear. Accordingly, the present-day Gulf attained its current form as a result of the separation of the Cape region from mainland and the break of the land barrier separating the two basins. Therefore, some present-day Gulf taxa, including Ligia, may have colonized and remained in the Gulf since northern proto-Gulf times. An early diversification (i.e., during proto-Gulf rather than modern Gulf times) may have contributed to the high levels of endemism observed in the northern Gulf [19].

Phylogeographic patterns of other taxa are also suggestive of colonization and isolation events during proto-Gulf times. Several fish lineages with disjunct distributions in the Pacific coast of Baja and the northern Gulf (i.e., not found in southern Gulf) show relatively deep Pacific-northern Gulf divergences. For example, divergence between Gillichthys mirabilis (with disjunct distribution) and its sister Gillichthys seta, endemic to the northern Gulf, is estimated to have occurred 4.6–11.6 Ma [63]. Similarly, Pacific/Gulf divergences for Leuresthes tenuis, Girella nigricans, and Hypsoblennius jenkins (other fish taxa with disjunct distributions) at the mitochondrial control region are 11.6%, 8.5%, and 7.9%, respectively [64]. According to a molecular clock of 0.85–2% per million years presumably specific to fish mitochondrial control region [52], some of the deeper divergences may have occurred during Miocene times (i.e., proto-Gulf times). Nevertheless, much shallower Pacific/northern Gulf control region divergences are reported for other fish taxa [0.6–2.3%, 64] with disjunct distributions, suggesting that more recent colonization events of the northern Gulf have also occurred.

As discussed above, our results suggest that Ligia colonized the Gulf during proto-Gulf times; and that the Cape region was the last portion of southern Baja to separate from the southern mainland. Interpretation of the sequence of events that led to the divergence of the Ligia Gulf clade from its sister lineage in the Pacific (yet to be identified or extinct), and to the subsequent divergence of the Gulf North and Gulf South clades is limited by the incomplete knowledge of the geological history. Below, we identify three possible vicariant scenarios that could explain the early history of the Ligia Gulf clade, although geological evidence for several aspects of these scenarios does not exist.

Careyes Clade

This clade, which represents another highly divergent lineage in our study area, is enigmatic because of its limited and peculiar distribution. It was found in only two localities (Puerto Vallarta and Careyes) separated by ∼94 Km of coastline; spanning a region that interrupts the distribution of one of the Gulf South clades (blue circles; Figs. 2 and 4). Gulf South lineages found south of the Careyes clade, form a monophyletic group (i.e., Manzanillo S20 to Barra Potosi S17) that represents a much younger lineage than the Careyes clade. How this Gulf clade ‘skipped’ the region occupied by the Careyes clade is intriguing. The divergence of the Careyes clade from the other lineages in our study may be related to the suggested formation of a Miocene (∼8 Ma) microbasin around Puerto Vallarta [46].

Concluding remarks, implications, and future work

As expected from its biology, extraordinary levels of allopatric genetic divergence among many localities of Ligia were observed in our study area. The spatial distribution of well-supported Ligia monophyletic groups corresponds strongly with geography, with little spatial overlap among clades; suggesting allopatric differentiation plays a major role in diversification of Ligia in this region. Extremely high mitochondrial genetic divergence is observed among the four main Ligia lineages identified, and within both, the Baja Pacific-Southern California and the Gulf clades. Although L. occidentalis is the only species usually recognized in the study area (in addition to L. pallasi in central California), high levels of genetic differentiation suggest the existence of multiple species. Indeed, based on our results and the type locality, only the Central California Clade would correspond to L. occidentalis. Therefore, the taxonomy of Ligia in the study area needs to be revised. This is important for conservation, as some divergent lineages have a very restricted distribution and, thus, are vulnerable to local anthropological pressures on the rocky intertidal.

Relationships among the four main lineages found in our study area could not be resolved. However, they are part of a well-supported clade that contains L. exotica, L. perkinsi, L. hawaiensis, and L. vitiensis; and excludes L. pallasi (distributed from central California to Alaska), as well as all other Ligia species examined. Inclusion of additional markers and Ligia lineages may help resolve these relationships.

Differentiation of many divergent regional clades of Ligia in our study area may be associated with past vicariant events. In contrast, shallow divergences within some of these clades across extensive distances may indicate recent exchange. For example, recent low sea levels may have facilitated recent exchange among localities in the upper Gulf and among localities within the mainland area south of the Gulf.

The phylogeographic patterns of Ligia in the Gulf of California and Baja Peninsula deviate from the pattern expected under a vicariant scenario often used by phylogeographers. This is surprising given that Ligia has characteristics of a taxon with a high potential for retaining phylogeographic signatures of past vicariant events that occurred in the study region (i.e., wide distribution, high allopatric differentiation, very divergent lineages, and extremely limited dispersal abilities). Although long-distance over-water dispersal cannot be completely ruled out as the cause of these conflicting patterns, the life history, exceptional degree of allopatric divergences, and the non-random phylogeographic pattern of Ligia, argue against this kind of dispersal as the cause of this discordance. Furthermore, some of the most “conflicting” phylogeographic patterns of Ligia appear to be concordant with aspects of alternative geological hypotheses that encompass an older origin of the Gulf. In this regard, it is surprising that the phylogeographic literature of this region conveys an impression that the geological history is well resolved and agreed upon, while this issue is subject to intense and ongoing debate among geologists. Congruence between the phylogeographic patterns of an organism with the characteristics of Ligia and an alternative geological hypothesis will contribute to this debate, and underscores that alternative hypotheses should be considered when examining phylogeographic patterns in this region. Examination of other taxa with restricted dispersal abilities, whose ancestors inhabited the region prior the formation of the Gulf of California, and whose current distribution includes key areas such as the southern Baja peninsula and the southern mainland, will allow for an adequate examination of alternative geological hypotheses. Shared patterns among multiple taxa with these characteristics should shed light on the complex history of this important biogeographic region. In addition, nuclear markers of Ligia should be examined to corroborate that the mitochondrial phylogeographic patterns extend to the rest of the genome.