2.5. Genetic differentiation
The pairwise FST coefficient obtained between the outbreak areas in Mm clearly indicated the separation of LU from the other two outbreak areas (LU vs. SM − 0.019, LU vs. OS − 0.022), while a low and non-significant (P > 0.05) value of FST coefficient was obtained between OS and SM (0.004). Similar results were found when considering the sex of the insect. Both males and females of Mm from LU showed significant values of FST (P < 0.05) coefficient between LU vs. SM and LU vs. OS. Taking into account the criterion of the sex of the insects, a statistically significant FST (FST = 0.011, (P < 0.05)) was also obtained for males between OS vs. SM.
In contrast to Mm, the second insect of interest, Mh, showed considerable differences in genetic differentiation patterns. The FST coefficient calculated for Mh between outbreak areas showed no statistical significance (P > 0.05) in all pairwise comparisons (LU vs. SM − 0.009, LU vs. OS − 0.004, OS vs. SM − 0.004), even when separation of sexes was taken into account.
Genetic differentiation between outbreak areas was reflected in more detailed differentiation between sampling sites. With respect to genetic differentiation among seven sampling sites representing mass occurrences of Mm, the pairwise FST coefficient showed high and significant values (P < 0.05) between LU1 vs. SM1, SM2, OS1, OS2 (from 0.018 to 0.034) and between LU2 vs. SM1, SM2, OS1, OS2 (from 0.028 to 0.036) (Table 4). In addition, when insect sex was considered, the FST coefficient generally mirrored statistically significant (P < 0.05) differences between the above sampling sites for LU1 in both sexes and for LU2 in males (Table S2). However, it is important to note that a greater number of statistically significant (P < 0.05) genetic distances were found for Mm males than for females (13 records vs. 6 records).
In turn, for Mh, the results of genetic differentiation between sampling sites showed only one statistically significant FST (P < 0.05) value (LU2 vs. SM2) (FST − 0.030, Table 4). The detailed results for males and females at the sampling site level are presented in Table S2.
Table 4
Pairwise FST values for Melolontha sp. for sampling sites (below diagonal M. melolontha, above diagonal, M. hippocastani). Statistically significant values for FST distance (P < 0.05) are marked in bold
Site | SM1 | SM2 | OS1 | OS2 | LU1 | LU2 | LU3 |
SM1 | | 0.008 | -0.006 | 0.01 | 0.001 | 0.008 | -0.003 |
SM2 | -0.004 | | 0.012 | 0.016 | 0.005 | 0.030 | 0.021 |
OS1 | 0.003 | -0.001 | | 0.004 | -0.003 | 0.009 | -0.008 |
OS2 | 0.010 | 0.005 | 0.004 | | -0.008 | -0.02 | -0.021 |
LU1 | 0.022 | 0.018 | 0.022 | 0.034 | | -0.008 | -0.015 |
LU2 | 0.036 | 0.034 | 0.028 | 0.036 | -0.007 | | -0.033 |
LU3 | 0.001 | 0.000 | -0.003 | 0.010 | 0.006 | 0.015 | |
The patterns of genetic differentiation for both closely related species were confirmed in the constructed PCoA biplot. The first two components of the principal coordinates analysis explained the vast majority (95.0%) of the total variation in genetic differentiation. The PCoA results showed a strong dominance of the first principal coordinate, which explained 92.8% of the variance and was associated with differences among species. The second principal coordinate was associated with differences within species and explained a further 2.2% of the variance. Visualization of the PCoA results showed a clear generic differentiation/distance between both Melolontha species and confirmed the large homogeneity within Mh sampling sites and the large separation of LU sampling sites with respect to sampling sites of SM and OS in Mm (Fig. 2).
2.8. Number of migrants
In 214 individuals of Mm and 207 of Mh, we detected a similar number of migrants per generation (11 and 9, respectively; P < 0.01), so that contemporary migration amounted to approximately 5% in both species. Males predominated among individuals assigned as migrants, accounting for seven male migrants in Mm and six in Mh.
In this study, we considered two important insect pest species, Mm and Mh, which are widely distributed in forest ecosystems across temperate Eurasia, to fill the knowledge gap on what genetic variation patterns ensure the success of these two species in dispersing in a changing environment. In examining genetic variation within the two species, we paid particular attention to genetic polymorphism and genetic structure (among and between populations). Consequently, it was possible to compare the differences in genetic parameters between these two species, especially that both were genotyped with the same set of microsatellite loci, the validity of which had been previously confirmed in population genetic studies of cockchafers (Tereba and Niemczyk 2017).
Our study showed that although these two species are closely related and were sampled in the same sites and time, there are clear differences in genetic polymorphism between Mm and Mh. This confirms recent results on the discrimination of these species based on single nucleotide polymorphism of mitochondrial DNA (Pedrazzini et al. 2021). Our estimates of genetic diversity parameters revealed higher mean numbers of alleles, allelic richness, HO and HE for Mm populations compared to Mh populations, providing a more detailed picture of heterozygosity for the studied species compared to the results of Masternak et al. (2022), in which both species had similar and moderate levels of heterozygosity. Yet the latter study was based on inter simple sequence repeat (ISSR) markers, where utility and polymorphism are lower than microsatellites.
Despite the high diversity in Mm, the departure from HW equilibrium and the ambiguous outcome of the bottleneck effect may suggest that this species has experienced an earlier population collapse or selection pressure. This could be partially explained by the slightly different ecology of Mm in comparison to Mh and the events both cockchafers have experienced in recent history. The occurrence of Mm is associated with both agricultural and forested areas, whereas Mh is generally restricted to the interior of forests, giving the impression of a more patchwork-like distribution (Billamboz 2014). Therefore, the latter species is likely to have lower genetic diversity, although distribution patterns should not be considered the sole reason for the results obtained. A brief literature review (including gray literature), as well as historical data from national archives (personal visit of MN to national archive in Tomaszów Maz., March, 2007), clearly show that cockchafers have been considered one of the most serious forest insect pests since the 19th century, i.e., since the introduction of the idea of planned forest management (Trąbczyński 1856; Satkowski 1899). It was not until the period of intensive use of pesticides in the 1950s and 1960s (DDT and HCH), especially in agriculture, that the number of cockchafers declined to the point where their protection was considered (Szczepanska 1972; Malinowski 1997). Considerations of a possible bottleneck in cockchafer populations associated with drastic control methods in the past and significant changes in agricultural practices primarily concern the common cockchafer, Mm, as it occurs in both agricultural and forested landscapes, and therefore may be more affected by human activities (Muska 2006; Zimmermann 2010; Woreta and Sukovata 2014; Woreta et al. 2016). Similar to what Juhel et al. (2019) suggest with respect to pollen beetle populations, in our study pesticides might have limited the dispersal of Mm populations to some degree, which influenced HW equilibrium and increased population genetic structuration (genetic differentiation). However, the results of the bottleneck effect analysis only partially support this hypothesis. Indeed, Garza & Williamson's indices demonstrate reductions in population sizes for both species in the past, while only some of the tests implemented in the Bottleneck software indicate such bottlenecks exclusively for Mm. These ambiguous results could be due either to limitations imposed by the available data (e.g., the small number of loci) or to differences in the calculation of bottlenecks (e.g., the first method might detect changes in a more distant past than the second method). The insecticides could cause a sign of a bottleneck in the more distant past, while recent massive outbreaks could override this earlier genetic evidence of a bottleneck. Thus, even if both Melolontha species have experienced reductions in population sizes in the past, these changes have had limited effects on current population sizes. Estimates of effective population sizes, although subject to wide credibility intervals, indicate large numbers of individuals contributing to further generations of the populations studied in the sampled areas. This evidence suggests that the cockchafers have not experienced an environmental disaster or other events detrimental to the species and, more importantly, that they do not appear to be driven by severe outbreak regimes.
Apart from recent historical events, the detection of some population genetic structuration in Mm (AMOVA, STRUCTURE, supported by the results of FST and PCoA) and a high degree of homogeneity within Mh populations may suggest that the two species studied underwent different processes in ancient times. However, the data regarding the phylogeography of these two important insect pests are largely lacking. The resent study by Pedrazzini et al. (2021) provides only limited insight into the phylogeographic diversity of Melolontha species by indicating very low intraspecific genetic diversity, which suggests a lack of spatial structuring. The mentioned authors examined only the Alpine populations of cockchafers with limited sampling effort, whereas nothing is known about the phylogeography of these species in the rest of their wide Palearctic ranges. Some genetic differences found in Mm in Poland and the absence of such a spatial structure in Mh may therefore indicate that the two species studied had different migratory pathways in earlier times. The Polish territory is known to be an important transient zone in the phylogeography of many Eurasian species, where phylogenetic lineages of different origins (from distinct glacial refugia) meet (Taberlet et al. 1998). This pattern is also known for many arboreal beetles (Kajtoch et al. 2022).
It is also important to note that the PCoA results showed a clear genetic distinction between the two Melolontha species. As the study by Giannoulis et al. ( 2011) showed, the marked chromosome difference between Mh and the other Melolontha species (Mm and M. pectorialis) indicates their ancient separation, with a likely interruption of the gene flow. A similar pattern of species distinction was recently found by Pedrazzini et al. (2021), based on CO1 mtDNA, showing closer genetic similarity of the Mm clade, which formed a sister cluster to the M. pectorialis instead of Mh, which formed separate cluster. This fits well with morphological differences, differences in polymorphism and the sympatry without hybridization between Mh and Mm (Giannoulis et al. 2011).
The vast majority of genetic diversity in Mm and Mh was found within populations (98.5% and 99.8%, respectively). A similar pattern of genetic diversity (with 82–86% of variation within populations) was found using mitochondrial sequences in another pest of trees - Scolytid Tomicus piniperda L. (Kerdelhué et al. 2002). Comparably large proportions of genetic variation within populations have been found in insect species whose ranges have recently expanded, which, as suggested by Conord et al. (2006), may be due to the increasing availability of a cultivated host. In spite of the fact that we cannot directly compare the behavioral patterns of other outbreak beetles with those of Melolontha spp. because the latter are polyphagous, it seems that current environmental changes are also promoting their development. Broadleaf tree species are almost exclusively the food base of cockchafer imagines. The ongoing changes in the species composition of forest ecosystems toward an increasing share of broadleaf tree species as a result of climate change and human activities are altering feeding habitats for cockchafers and other organisms. In addition, a longer growing season, higher CO2 concentration and nitrogen deposition are accelerating the dynamics of forest stands (Lindner et al. 2010; Socha et al. 2021), which together is leading to a significant increase in the quantity and quality of the food base for cockchafers and, consequently, their fecundity. Recent studies have shown that when adult cockchafers have a good foraging habitat in a forest ecosystem, they do not fly long distances to search for oviposition sites and usually lay their eggs in close proximity to broadleaf trees (Niemczyk 2015; Cours et al. 2021). The choice of oviposition site is an important aspect of female behavior, as it has a major impact on the fitness of a species (Cury et al. 2019). Estimates of larval populations of Melolontha spp. of up to 320,000 individuals per hectare in their third larval stage (the oldest) in sampled areas (Niemczyk 2015; Niemczyk et al. 2019) indeed indicate good foraging habitats. Importantly, with such large populations, alleles are unlikely to be lost through genetic drift. This is confirmed in our current study by the relatively high genetic polymorphism and mostly infinitive estimates of effective population sizes of the two cockchafers studied, but particularly that of Mm.
In such large populations, the mating system is random, suggesting panmixia. A number of migrants per generation greater than 1 allows gene flow and maintains equilibrium in the population (Varvio et al. 1986). Our estimated migration rates are 10-fold greater than this threshold for both species and the study by Masterak et al. (2022) also shows substantial exchange of migrants between populations per generation. Importantly, our study also revealed differences in the number of migrants between females and males in favor of the latter. This raises important questions about the migratory behavior of females and males, whether there is a trade-off between migration and reproduction, and whether males and females bear the differential costs of migration (Rankin and Burchsted 1992). During their above-ground life, adult female cockchafers undergo two or three ovarian cycles, each characterized by purposeful migrations that lead them to feeding sites and from there back to egg-laying sites (Stengel 1974), which forces philopatry. This activity is associated with maturation of the eggs and is accompanied by changes in the neuroendocrine system (Stengel 1974). The main characteristic of this migratory behavior is the reversal of the sense of direction of flight in females that are ready to lay eggs. Like females, males fly toward feeding sites after leaving the soil and are able to resume their original orientation when released in an unfamiliar area. They usually move within the adult feeding area where they feed and mate. Unlike the females, however, the males never show a reversal of flight direction. Such different migratory behavior between males and females may explain more frequent long-distance migratory flights among cockchafer males. Another aspect of the phenomenon related to the differences in the number of migrants between males and females may arise from the fact that flight seems to be associated with some reproductive costs in females, which has been demonstrated in other insect species and broadly discussed by Rankin and Burchsted (Rankin and Burchsted 1992).
In our study, conducted at a fine spatial scale, we were unable to demonstrate clear genetic-geographic correlation distances for either species studied. Although we obtained a statistically significant result for females of Mh (p < 0.05) when we performed site-level correlations separately for males and females, supporting the hypothesis of greater philopatry of females and stronger migratory behavior in males, this assumption failed in Mm. In the case of the latter species, we obtained a statistically significant result for males, which is difficult to interpret. Perhaps a greater number of sites from geographically more distant areas would be required to nuance the genetic-geographic distances.