Proximity to the introduction site and local dynamics
Our admixture analysis with the set of species-specific loci showed bidirectional introgressive hybridization between I. elegans andI. graellsii . Individuals from both hybrid regions, north-west and north-central Spain, mainly belonged to the introgressed I. elegans and introgressed I. graellsii classes, and less frequently to the backcross to I. elegans , backcross to I. graellsii , F1 and F2 hybrid classes. Almost no pure I. elegans or pure I. graellsii were found in the hybrid regions.
When comparing the distribution of parental and hybrid classes between north-west and north-central hybrid regions, all classes were at similar frequency proportions except for the introgressed I. elegans class, which was significantly higher in the north-west hybrid region. The proximity to the introduction or source locality could affect hybridization outcomes (Fitzpatrick et al., 2010; Lepais et al., 2009), as the rate and directionality of introgression can be influenced by the species’ relative abundance and therefore by gene flow from the native to the non-native species. A non-native colonising species is usually rare and matings with the native species are likely, as was detected in other studied organisms (Fitzpatrick et al., 2010; Lepais et al., 2009; Quilodrán, Nussberger, Montoya-Burgos, & Currat, 2019). Introgressed genes can reach high frequency in the non-native species by a rapid demographic growth, resulting in asymmetric introgression of neutral genes (Currat, Ruedi, Petit, & Excoffier, 2008). The north-central hybrid region is closer to the allopatric distribution of I. elegans and disconnected to theI. elegans populations from the north-west hybrid region. In fact, the number of populations dominated by I. elegans , and the abundance of I. elegans in populations with both species in the north-western hybrid region is lower than in the north-central hybrid region which is closer to the allopatric distribution of I. elegans (Boudot et al., 2009; Sanchez-Guillen, Van Gossum, & Cordero Rivera, 2005; Sánchez-Guillén et al., 2011)
In both hybrid regions local populations showed pronounced variation in species composition and in the degree of hybridization and introgression, which suggests that the local dynamics is even more distinct than the regional dynamics in this system (Fig. 2A-B). In four out of the six study populations where both species co-occur at different numbers (Cañas, Louro, Perdiguero and Villar), nine individuals morphologically classified I. graellsii were genetically assigned to F1 or F2 hybrid classes. This is consistent with observations in the laboratory that F1 hybrids are morphologically similar to I. graellsii (Sanchez-Guillen et al., 2005), and independent of the reciprocal cross direction (I. elegans male mated with I. graellsii or hybrid female, or I. graellsii male mated withI. elegans or hybrid female) that has produced the F1 or the F2 hybrid (RAS-G personal observation). Local dynamics was dominated by periodic dry-ups followed by recolonizations by one or both species. For instance, Cañas and Villar (north-central populations) were emptied and recolonized a couple of years before we sampled them. Similarly, the north-western population Louro became extinct due to salobrisation by seawater in 2009, and was recolonized by both species, forming a hybrid population with high levels of admixture. In Arreo and Valbornedo we did not detect F1 or F2 hybrids, even though both species are present. Even the six studied populations with only one of the genotypical-parental species [Doniños, Laxe and Menorca (onlyI. elegans ), and Mateo, Valpierre and Xuño (only I. graellsii )] showed different levels of introgression or back crosses, indicating that although only one species was detected in the sampling years both species were possibly present in the population in the years before. This can be the case of Xuño that had both species during 2003-2006 and was naturally dried in 2007 and after that recolonized byI. graellsii , which explains the observed levels of admixture.