Sympatric speciation in the Simulium arcticum s. l. complex (Diptera: Simuliidae): The Rothfels model updated

Abstract We tested the Rothfels sympatric speciation model for black flies by comparing all available data for sex‐chromosome diversity with the geographic locations of larval collection sites within the Simulium arcticum complex of black flies (Diptera: Simuliidae). Five separate data sets equaling about 20,000 larvae were included from throughout the geographic range of this complex. We record a total of 31 taxa having unique sex chromosomes, all of which demonstrate linkage disequilibrium with most taxa sharing autosomal polymorphisms. All siblings share portions of their distributions with S. negativum, the presumed oldest member of the complex. Twenty‐one of 22 cytotypes have distributions within the ranges of siblings thus supporting the sympatric speciation model of Rothfels. Chromosomally diverse sites may require analysis of as many as 200 larvae to be properly described. There is no effect of any inversions influencing the occurrence of other inversions. Finally, we report a new cytotype, Simulium arcticum IIL‐6, which we originally discovered in Alaska. Aspects of future genomic research are discussed as they relate to the main chromosomal structural/functional tenants of the model. OPEN RESEARCH BADGE This article has earned an Open Data Badge for making publicly available the digitally‐shareable data necessary to reproduce the reported results. The data are available at https://doi.org/10.6084/m9.figshare.7719398

morabine grasshoppers and by Bush (1969) and Bush et al., 1989) on the Rhagoletis genus of fruit flies have kept the topic relevant.
Moreover, there has been a recent resurgence in the speciation debate (Ayala et al., 2011;Feder & Nosil, 2009;Guerrero, Rousset, & Kirkpatrick, 2012;Hoffman & Riesberg, 2008) and Gavrilets (2014) summarizing what we now know about speciation. The works of White and Bush were based on chromosome change in species of insects, yet in their monumental coverage of speciation Coyne and Orr (2004) state that, "It is far from clear whether chromosomal speciation is common in animals generally. Indeed, we know of no compelling evidence for chromosomal speciation (and by extension, sympatric speciation) in animals." Doellman et al. (2018) have used molecular approaches in Rhagoletis to document host related variation in divergent life history adaptation through increased selection on diapause and reduced gene flow. They suggest host races may be recognized as different genotypic entities and perhaps even good species in sympatry.
Thus, the argument may not be whether sympatric speciation occurs or is even possible but rather if it is widespread. Interpretations of allopatric speciation are clearly both common and widespread in many groups of organisms (Coyne & Orr, 2004;Price, 2008).
However, extant allopatric distributions neither confirm nor deny allopatric origins. This is because allopatric distributions make tests of reproductive isolation difficult, if not impossible, by virtue of geographic separation and lack of opportunity to mate. It may be possible for different types of organisms to become reproductively isolated in different ways (Feder et al., 2017). A case in point is the black fly, Simulium ruficorne s. l. with its present-day allopatric distribution of cytotypes/siblings (Cherairia & Adler, 2018). Rothfels (1989) proposed a sympatric speciation model for black flies (Simuliidae) based on the somewhat fragmentary data available at that time. Larval black flies possess enlarged polytene chromosomes (Landau, 1962;Rothfels & Nambiar, 1981) and chromosomal rearrangements are relatively easy to describe (Adler, Currie, & Wood, 2004;Rothfels, 1979;Shields & Procunier, 1982). The Rothfels model (1989) was based on: (a) "frequent and often exclusive involvement of changes in the sex-chromosome systems," (b) "sex-chromosome polymorphisms that may display linkage disequilibria," (c) "complexes that differ only in sex chromosomes and share extensive ancestral autosomal polymorphisms," (d) "sympatric or widely overlapping distributions of the most closely related species" (taxa), and (e) "species that differ in their biology and perhaps their present-day distributions." Of interest is that one of the tenants of the Rothfels model pertaining to chromosomal structural rearrangements and their redistribution is still observed; namely, one and the same inversion/band can be autosomally polymorphic, sex linked, fixed and/or lost within various taxa of the group.
Evolutionary biologists are left with the challenge of analyzing large numbers of genetically identified individuals throughout the current distributional ranges of closely related taxa for clues that might suggest support for either an allopatric or a sympatric origin. In the present context, dispersal distances are important since fidelity to natal sites could allow in situ genetic change. Nothing is known about dispersal distances in members of the Simulium arcticum complex but Shields, Christiaens, Luvan, and Hartman (2009) have shown that genotypes of taxa of the S. arcticum complex at the Clearwater River in western Montana are essentially identical over a three-year period. Dispersal distances of less than 15 km seem typical of most species of black flies (Baldwin, West, & Comery, 1975;Bennett, 1963;Moore & Noblet, 1974). Critical in the present context is that female black flies generally do not undergo longdistance dispersals after emergence (Adler et al., 2004) although it has been shown that river corridor affects chromosome diversity (Shields & Hokit, 2016).
At least 11 species complexes (cytospecies within a single morphospecies) of black flies have been described for North America alone (Adler et al., 2004;Rothfels & Featherston, 1981) and the number increases to 45 complexes world-wide (Adler, 2019; http:// biomia.sites.clems on.edu/pdfs/black flyin vento ry.pdf). Moreover, without the original cytogenetic analyses, these complexes of reproductively isolated sibling species might have gone unrecognized (Rothfels, 1979(Rothfels, , 1989. The S. arcticum Malloch complex of North America is one of the most diverse in existence, second only to the S. damnosum complex in Africa (Conflitti, Shields, Murphy, & Currie, 2015b;Shields, 2013). Five sibling species were initially described within the complex (Procunier, 1984;Shields & Procunier, 1982). Four of these (S. brevicercum, S. saxosum, S. arcticum s. s., and S. negativum) were later studied in detail and given full species status by Adler et al. (2004), Simulium arcticum IIL-1 has not been studied further. Comparisons of taxa using mitochondrial and nuclear DNAs show that chromosome inversions occur in the initial stages of differentiation when no morphological changes have yet occurred (Conflitti, Shields, Murphy, & Currie, 2015a, 2016. Further, taxa within complexes usually form a continuum from presumably relatively recent cytotypes to full sibling species. Cytotypes are defined as taxa having unique sex-linked inversions but whose reproductive status is yet undetermined (Adler et al., 2004;Shields, 2013). Nine taxa are given species status and the other 22 have: (a) unique paracentric chromosomal inversions linked to sex, (b) are not monophyletic in phylogenetic trees based on comparative sequences of DNAs (Conflitti, Kratochvil, Spironello, Shields, & Currie, 2010;Conflitti, Shields, & Currie, 2012;Conflitti, Shields, Murphy, & Currie, 2014), and (c) are assumed to be in the early stages of reproductive isolation (Shields, 2013). S. negativum is both morphologically unique and monophyletic in our DNA trees (Conflitti, Shields, Murphy, & Currie, 2016). Moreover, it is the oldest extant member of the complex among those analyzed molecularly. Its separation from the remainder of the complex is estimated at 467,500 YBP (Conflitti et al., 2016) and its putative sex-determining gene or genes is (are) located in the long arm of chromosome I, unlike all other members of the complex (Adler et al., 2004;Shields & Procunier, 1982). Compelling is the original observations of Rothfels (1989) and to a greater extent our own large data set on Simulium arcticum that suggest that newly discovered taxa arise almost exclusively within the geographic distributions of other taxa of the complex whose own distributions are large, suggesting a sympatric origin or at minimum arguing against separation by distance before reproductive isolation.
The Rothfels model has not been rigorously tested on a taxon complex throughout its range of distribution. Consequently, we chose to revisit the model using the available data on the S. arcticum complex. Herein, we first use the criteria for sympatric speciation originally set forward by Rothfels (1989) on our enlarged data set for the S. arcticum complex to revisit the sympatric model. Importantly, we compare sex-chromosome diversity of all known members of the S. arcticum complex from 13 states and six Canadian provinces throughout the distributional range of the complex in North America. The final prediction of the Rothfels model suggested that species should differ in their biology and possibly geographic locations. By using canonical correspondence analysis, we have previously shown that despite significant overlap, all siblings of S. arcticum are ecologically unique (Conflitti et al., 2015b). Further, cytotypes are either ecologically unique, are associated with one another, or with particular siblings (Conflitti et al., 2015b). Thus, for this group, ecological and chromosomal differences develop early in lineage formation, suggesting that local adaptation may be involved in diversification (Conflitti et al., 2015b;Pramual, Kuvangkadilok, Jitklang, Tangkawanit, & Adler, 2012;Shields & Kratochvil, 2011).
Secondly, we ask how much analysis is sufficient to accurately describe sex-chromosome diversity at any collection location. We have analyzed the chromosomes of more than 1,000 individuals from spring collections at four different sites (Shields et al., 2009) and thus were able to complete this analysis here. Thirdly, because nine sibling species and 22 cytotypes have been described on the basis of unique sex-chromosome morphologies for the S. arcticum complex, we questioned whether the break points of one inversion might influence the break points of other inversions. Finally, we report a new cytotype, S. arcticum IIL-6, which we originally discovered in Alaska.
Scientific justification for this research rests on the facts that: (a) a return to the sympatric speciation question originally proposed by Rothfels (1989) seemed warranted, (b) this is the largest data set assembled for a species of black fly in North America, (c) all data are chromosomally verified, and (d) the complex has been sampled throughout its distributional range.

| Changes in sex-chromosome systems
We simply list taxa whose sex chromosomes differ. In almost all cases, differences in sex chromosomes are absolute (the exception is S. brevicecum whose chromosomes are standard (noninverted) in both sexes). For example, in a 3/14/2006 collection of S. arcticum from Rock Creek, Missoula County, Montana all females (n = 447) had standard, noninverted, chromosomes while F I G U R E 1 The IS-1 autosomal inversion in the short arm of chromosome I. Numbers within brackets indicate regions of the entire arm of the chromosome from 1 to 19. Note that the IS-1 inversion encompasses nearly the central 1/3 of the chromosome and forms a reverse loop (polytene chromosome pairing) from region 6 to region 12. Dark arrows indicate regions of obvious homology all 261 male larvae were either IIL-9 st/i (n = 106) or IIL-19 st/i (n = 155; Shields et al., 2006).

| Linkage to sex
For each taxon, we determined whether paracentric chromosomal inversions were linked to either sex. This was possible since for all analyses, the gonads of each larva were placed on the same microscope slide as the polytene chromosomes. A chromosome inversion was judged to be sex-linked if it was found exclusively, or very nearly so, in one sex or the other. Chromosomal rearrangements were judged to be autosomal if they occurred in each sex equally or nearly so.

| Sharing of autosomal polymorphisms
We calculated the proportion of individuals which were heterozygous for the most common autosomal polymorphisms using data sets for which autosomal polymorphisms were chromosomally described or published ( Figure 1).

| Sex-chromosome diversity vs. geographic location
For the entire data set, we constructed a database listing the collection site, its geographic coordinates when available, the number of S. arcticum taxa present and the number of larvae for each taxon when available. From these data, we could determine the extent of the geographic distributions of all taxa of S. arcticum. A key issue was to describe the entire geographic range of each taxon to determine whether there was overlap with other taxa or not. Distributions within geographic ranges might suggest, but do not prove, a sympatric origin. Mutually exclusive distributions might suggest, but do not prove, an allopatric origin.
Locations for 307 sites were determined with varying accuracy.
There were 96 sites associated with recent collections using GPS that have an accuracy of plus or minus 5 m. There were 119 legacy sites with good location descriptions that we are confident are within 1 km of the collection location. Finally, 92 sites had poor location descriptions often listing only a county. We used the centroid of the county location for these sites resulting in a spatial accuracy of approximately 100 km. Locations for each chromosome taxon along with associated sibling species and cytotypes were used to create an attribute table.
Following a process similar to the methods used by Swenson and Howard (2005) attribute, data were mapped and analyzed using ArcGIS Pro 2.2. First, point features were created for each sibling species and cytotype. The geographic extent of each species and cytotype was defined using the minimum bounding geometry tool with the convex hull option to minimize the assumed geographic extent of each species/ cytotype. It was not possible to use the minimum bounding geometry tool for cytotypes known from fewer than three locations. For these cytotypes, point and line features were created using the buffer tool and a buffer distance of 5 km. This is well within the known dispersal distance of individuals of this species complex (Adler et al., 2004). For all species and cytotypes, the resulting polygons were clipped using a mask of North America. Then, the union tool was used to measure the contact between range extents of each pairwise combination of siblings/cytotypes. Finally, the number of contacts was quantified for each and categorized as a contact with a sibling species or with a cytotype.

| How much collection is sufficient?
We had previously analyzed spring larvae from four locations (Blackfoot River, n = 774; Little Prickly Pear Creek, n = 1,330; Little Blackfoot River, n = 1,359 and the Clearwater River, n = 2,197) in western Montana to determine reproductive status and continuity of chromosome taxa from year-to-year (Shields, 2013;Shields et al., 2009). Given these large sample sizes, we were able to ask the question, how much analysis must be done to fully characterize the black fly cytogenetic diversity at any one site in spring?
To assess this statistic, we randomly chose 100 individuals from each collection and determined its taxonomic diversity. We then increased the random sample number to 200, 300, 400, etc. to determine at what sample size the taxonomic diversity failed to increase markedly. We are aware that some taxa of black flies are multivoltine (have more than one generation) so we generally restricted our analyses only to spring, first emergent, larvae.

| Do different sex-linked inversions have at least one breakpoint in common?
More than 80 unique paracentric chromosomal inversions occur within the S. arcticum complex, of which 30 are sex-linked (Shields, 2013;Shields, unpublished). Given one sex-linked inversion, one wonders whether the probability of a second sex-linked inversion might be increased if the two inversions have one or the other break points in common. To assess this, we compared all break points of sex-linked inversions (30 × 30 × 2 = 1,800) to all other break points of other sex-linked inversions within the complex. If either of the two break points of a sex-linked inversion was identical to that of another break point of another inversion the taxa compared were given a plus (+), if not, they were given a minus (−).
This cytotype was discovered in the process of analysis of new sites and new drainages in Alaska. It was given cytotypic recognition because the sexes absolutely differed regarding sex chromosomes.

| Changes in sex-chromosome systems
There is abundant evidence within the S. arcticum complex for changes in sex-chromosome systems (Table 1). With extensive study throughout the range of distribution of S. arcticum, there are at least 31 chromosome taxa whose sex-chromosomes differ. These unique sex chromosomes are the only indication that these taxa differ with the exception of S. negativum, which can be morphologically identified (Adler et al., 2004;Shields, 2013).

| Inversions linked to sex
The majority of taxa within the S. arcticum complex display complete linkage to sex and thus linkage disequilibrium (Table 2). In fact, of the seven taxa that do not display complete linkage to sex, none has a linkage to the Y chromosome of less than 0.967 (Table 2).

| Sharing of autosomal polymorphisms
It is abundantly clear that taxa within the S. arcticum complex share autosomal polymorphisms (Table 3). Here, we report the proportions of heterozygotes among some S. arcticum taxa. The autosomal polymorphism, IS-1, occurs in 13 of 15 taxa, while the autosomal inversion, IL-1, occurs in 11 of 15 taxa. TA B L E 1 Taxa of the Simulium arcticum complex. Cytospecies (full sibling species)

| What is the appropriate sample size?
For spring collections, our data suggest that detected diversity of sex chromosomes reaches about 90% after about 200 samples are analyzed and that detected diversity increases very slowly after that (Figures 2-4). It is important to acknowledge that our comparisons are all from spring collections and presumably from first generation larvae.

| Chromosomal break points
Only 46 of 1800 (2.6%) possible break points were in common (Table 5A-C). Moreover, of the 46 break points in common with others, there appears to be no obvious break point "hot spot." This may imply that the breakpoint of any inversion has little to no effect on subsequent break points of inversions in the complex.

| The IIL-6 cytotype
Of 206 larvae analyzed from the Delta Clearwater River and Monument Creek sites in Alaska, all males (n = 113) were heterozygous for the IIL-6 paracentric inversion while all females (n = 93) possessed the standard (noninverted) sequence (Table 6 and Figure 5).
This sex-linked chromosomal sequence inverts the entire section 56 of the long arm of chromosome II ( Figure 5). Table 7 lists the most common autosomal polymorphisms in populations of IIL-6. Rothfels (1989)

| Linkage disequilibrium
The Rothfels model also predicted that changes in sex-chromosome systems would result in linkage disequilibrium between the sex chromosomes. Linkage to sex, particularly in males, is almost absolute in taxa of the S. arcticum complex. In the majority of types, linkage to sex is 100% and in those that are not, linkage is only slightly less than 100%. Thus, the second criterion of the Rothfels model is fulfilled.
This is a characteristic common in many other species complexes of black flies (Adler et al., 2004).

| Autosomal polymorphisms
The were also heterozygous for the IS-1 inversion (Shields et al., 2006).
Thus, there may be two types of sex chromosomes at the Blackfoot River in mid-May.
We have suggested via molecular analysis that gene flow occurs between siblings of black flies (Conflitti et al., 2016).

| How much analysis is necessary?
Sex-chromosome diversity at sites extensively analyzed rose dramatically up to a sample size of 200. After that, diversity rose slowly and only about 10% diversity was added if 600 larvae were analyzed. This suggests that most black fly sites (even in this study) are insufficiently analyzed.
Chromosomally diverse sites will require more analysis than taxon-pure sites. For example, the Little Blackfoot River at Elliston, Montana is one of the most diverse sites we know of with two sibling species and nine cytotypes present. In order to detect all of this diversity, we had to analyze nearly 800 individuals on 4 April 2011 The number of different Y chromosomes plotted against the log of the sample size F I G U R E 2 The statistic indicated is the standard error. The average number of chromosome taxa with overlapping (contact) ranges is grouped as: (1) siblings with other siblings, (2) siblings with cytotypes, (3) cytotypes with siblings, and (4) cytotypes with other cytotypes. Note that because siblings have much larger ranges than cytotypes, the average number of contacts that siblings have with cytotype is larger than the number of contacts cytotypes have with siblings (Shields, 2013

| Break points
The basal region of the long arm chromosome II may have "hot spots" for sex-linked inversions. Given this, we questioned why this region had so many sex-linked inversions. By extension, does a one inversion breakpoint increase the probability of another breakpoint? Only 46 of a possible 1800 (2.55%) break points are in common. This suggests that one breakpoint may not influence another breakpoint.
The 46 break points in common among taxa may be an overestimate because it is difficult to determine specifically the exact breakpoint in regions between obvious chromosome bands. Some regions of the long arm of chromosome II are so-called "puffing regions" and since there are few specific and detailed markers in these regions, our analyses might be inaccurate. If break points appeared similar in these regions they were scored as positives. Molecular studies of the basal portion of the IIL region might reveal why so many inver-

| The IIL-6 cytotype
This cytotype was found at only two sites, the Delta Clearwater River and Monument Creek of the Fairbanks-North Star Borough, Alaska (Table 5). These sites are only 97.37 km (60.5 miles) apart. It is possible for gravid females of at least some species of black flies disperse this distance (Adler et al., 2004). Figure

| Caveats about the data
Although we claim to have analyzed S. arcticum larvae throughout the general range of distribution, we have not analyzed larvae from every drainage. It is likely that S. arcticum can be found in additional unstudied drainages. The large majority of samples (72%) are precisely located. For those that are not, we used the centroid of the county resulting in a special accuracy of from 10 to 100 km. This does not affect the outcome of the analysis since the overall distribution of the species complex covers tens of thousands of square miles.
Second, the current distributions of taxa within the S. arcticum complex may give little, if any indication, of the process or processes by which these taxa have become reproductively isolated. It is likely that taxa have undergone additional evolutionary processes since they originally arose. Rothfels (1989) argues that continentally distributed taxa may be difficult to interpret as to type of origin since sympatry was the reason they were detected in the first place.
The presence of sperm in the male gonad indicated that the Y chromosome was male determining as in other black flies (Rothfels & Nambiar, 1981) and unlike Drosophila. (C)   IIL-3  IIL-15  IIL-73  IIL-6  IIL-11  IIL-21  IIL-7  IIL-12  IIL-13 IIL-73   (Henderson, 1986a(Henderson, , 1986bNewman, 1983) and S. tuberosum in eastern North America (Landau, 1962;Mason, 1982Mason, , 1984  The original morphospecies of classical black fly taxonomy almost always becomes a complex of sibling species when detailed cytogenetic analyses are conducted. As suggested in this study, types might arise from preexisting types through chromosome rearrangements in sympatry. This may be a necessary condition for sympatric speciation.

| SUMMARY
All of the criteria set forward by Rothfels in his sympatric speciation model have been fulfilled by additional sampling and analysis. What may never be known is whether present-day distributions of taxa within the complex reflect origins. However, the weight of evidence suggests that taxa of the S. arcticum complex may give rise to new taxa via a sympatric model. Our observations also suggest that extensive sampling and analysis are necessary to adequately characterize the chromosome diversity at any one site and that break points of inversions seem not to influence the origin of additional inversions.
Future genomic 3D analysis (Adler, Hamada, Nascimento, & Grillet, 2017) of inversion hotspots within the complex should prove invaluable for defining the potential structural integrity/molecular makeup of inversions and their potential functions.
One implication of these results is that chromosomal change (including inversions) and the consequent lack of recombination within inverted regions may give rise to sex-chromosome complexes that subsequently diverge. We emphasize that our results relate only to black flies of the S. arcticum complex. A similar study may be difficult in organisms other than black flies since many of those organisms do not possess polytene chromosomes and the detail seen here may not be possible.  (Conflitti et al., 2016). We thank Dr. Elmer Gray (Univ. of Georgia) for reviewing an earlier draft of this manuscript and especially Dr. Peter Adler, Department of Entomology, Clemson University, for inclusion of his data on S. arcticum (Adler et al., 2004), for his review of an earlier draft of this manuscript, and for encouragement, continued interest and support of our work. We thank three anonymous reviewers and the associate editor for Ecology and Evolution.

ACK N OWLED G M ENTS
Finally, the late, Klaus H. Rothfels, is remembered for initially starting our interest in black flies.

CO N FLI C T O F I NTE R E S T
None declared. Note: S. arcticum IIL-6 is also characterized by meiotic development in mature male larvae to sperm and the absence of B chromosomes.

AUTH O R S CO NTR I B UTI O N S
GFS conceived the idea, collected most of the material, assembled and analyzed the data, wrote the original manuscript and responded to reviewers. WSP collected and analyzed some of the data and contributed to revising the manuscript.

DATA ACCE SS I B I LIT Y
Original data based on sex-chromosome type versus geographic locations of collections are archived in Dryad, a publicly accessible repository under the accession number (https ://doi.org/10.6084/ m9.figsh are.7719398).