Discussion
Impacts of commercially produced B. terrestris are more complex in the Mediterranean region where B. t. dalmatinus is native subspecies and large numbers of commercial colonies have been used continuously for 25 years. In this study we divided the sampled populations into three groups; 1) greenhouse (commercial) populations 2) nearby greenhouses and 3) far from the greenhouses populations. We examined the genetic structure of different commercial stocks and native populations and assessed the degree of genetic introgression into native populations using twenty microsatellite markers and two mitochondrial genes. Our results demonstrate that the values ofFST between populations were not significant, and the maximum FST value was only 0.045 (Table 3). Therefore, commercial and native populations were not differentiable based on twenty microsatellite markers. Similarly, structure analyses supported this pattern. According to the best K value (K =4), the populations are separated into four distinct clusters. However, there are transitions between these four main clusters and it has been determined that populations could not been clearly separated from each other. Neighbor-Joining (NJ) trees inferred from the mtDNACOI (A) and cyt b (B) gene regions could also distinguish groups clearly (Fig. 3). A possible reason for the observed low genetic differentiation between commercial and wild bumblebee populations is that although the origins of the commercial colonies were not well known, commercial and native B. terrestris populations probably derived from a common ancestor. A second explanation is that possible selection process in commercial breeding strains may have not significant impacts on the genetic differentiation between the two types of populations. Similarly, a few previous studies of B. terrestris showed that the mainland European B. terrestrispopulations had a relatively homogenous genetic structure but had clear differences for some island subspecies (Estoup, Solignac, Cornuet, Goudet, & Scholl, 1996; Lecocq et al., 2013; Widmer, Schmid‐Hempel, Estoup, & Scholls, 1998). But a recent study on B. terrestrisshowed that Irish B. terrestris populations were highly differentiated from British and continental B. terrestrispopulations, the latter two showing higher levels of admixture (Moreira, Horgan, Murray, & Kakouli-Duarte, 2015). In addition, Cejas, Ornosa, Munoz, & De la Rua (2018) recently found an effective marker (16S ) to differentiate B. t. terrestris and B. t. lusitanicus populations (endemic in Iberian Peninsula).
Native B. t. dalmatinus populations show considerable differences in the life cycle patterns depending on environmental conditions. Along the some coastal area of the Mediterranean region, native B. t. dalmatinus populations are active from November until July. New queens enter summer aestivation and remain dormant during the dry season until the rainy period starts (September-November) (Gurel, Gosterit, & Eren, 2008). Therefore, life-cycle patterns of aestivated native B. t.dalmatinus populations are synchronized with the greenhouses tomatoes production period (September-July) in coastal Mediterranean region where the commercial colonies have been widely used. However, aestivated native B. t. dalmatinus populations have become extinct or rare in Mediterranean coastal areas because of both introgressions of commercial stocks and habitat destruction, and heavy urbanization. Some authors have also implied that commercialization processes are capable of changing some colony traits of native B. terrestris colonies (Gosterit & Baskar, 2016; Ings, Ward, & Chittka, 2006). The genetic introgression could lead to the loss of specific traits and local adaptations (Evans, 2017). In this study, within the group of the more distant populations FS is the only aestivated population, located more than 30 km away from greenhouses areas and occurs in the one of the best-preserved ancient cities in the Mediterranean region. In our study, although the values ofFST between populations were not significant, according to PCA results, FS population separated from other populations. Therefore, we suggest that FS population should be initially conserved to avoid extinction.
In the mitochondrial analysis, two haplotypes (haplotype A; 28.8 % samples and haplotype B; 71.2 % samples) were found for the COIgene. While haplotype A was common in commercial populations, haplotype B was common in both nearby greenhouses populations and far from the greenhouses populations. In our study, six haplotypes (H1-H6) were found for the cyt b gene. Considering haplotypes and their frequencies, we found that the most abundant haplotype was H4, which was detected in 43 individuals (% 63.2). While haplotypes 4 and 5 are known to belong to the B. t. dalmatinus, haplotypes 1, 2, 3 and 6 are known to belong to both B. t. dalmatinus and B. t. terrestris.These haplotypes were found in the commercial populations and nearby greenhouses populations. The detection of haplotypes 1, 2, 3 and 6 in the commercial populations and nearby greenhouses populations potentially suggests that commercial producers in Turkey either have also been producing and selling non -native B. t. terrestris in Mediterranean region where B. t. dalmatinus is native subspecies or mixing B. t. terrestris and B. t. dalmatinus in order to set up breeding stocks. We don’t know whether B. t. terrestrishas become established in the Mediterranean region where it is not native. Some environmental factors can prevent its establishment in this region. There is some evidence to support this. Lecocq, Rasmont, Harpke & Schweiger (2016) considered that B. t. terrestris would be able to survive as active colonies in south of South America, New Zealand, southeast Australia, and Tasmania based on 19 bioclimatic variables and implied that Mediterranean region is not climatically suitable area for B. t. terrestris. On the other hand,because of high demand from consumers for a year-round supply of greenhouse production, highland (at different altitudes and climatic conditions) greenhouse tomato production in Turkey has been increasing in recent years. Therefore, commercial B. terrestris colonies are used for pollination all year round. We suppose that B. t. terrestris is able to establish in the wild particularly adjacent to the highland greenhouse areas and spread.
It will be useful for future ecological studies to determine locally geographic areas where the establishment success of commercial queens in greenhouses could be higher. In addition, because of morphological similarities and difficulties, new techniques and molecular markers to separate commercial strains, subspecies, and ecotypes and to determine the introgressions are needed. As mentioned in previous studies (Evans, 2017; Goka, 2010; Ings, Ward, & Chittka, 2006; Koide, Yamada, Yabe, & Yamashita, 2008), some special precautions; such as using of mesh to cover glasshouse vents, using of queen excluders to keep gynes in the nest and disposing of old nests to restrict the escape of males and gynes must be taken to prevent both novel subspecies or commercial stocks introductions and the spread of pathogens and pests. Despite some restrictions, regulation of the movement of B. terrestrissubspecies is problematic and ineffective. We still have little information on the origin of commercial stocks. To reduce the risk of non native subspecies and genetically different populations and the risk of spreading pathogens and parasites, effective control methods would seem to be necessary even without the establishment. Furthermore, there is an urgent need to develop conservation strategies for native bumble bee populations.