Cryptic Biodiversity and the Origins of Pest Status Revealed in the Macrogenome of Simulium colombaschense (Diptera: Simuliidae), History’s Most Destructive Black Fly

The European black fly Simulium (Simulium) colombaschense (Scopoli), once responsible for as many as 22,000 livestock deaths per year, is chromosomally mapped, permitting its evolutionary relationships and pest drivers to be inferred. The species is 12 fixed inversions removed from the standard sequence of the subgenus Simulium. Three of these fixed inversions, 38 autosomal polymorphisms, and a complex set of 12 X and 6 Y chromosomes in 29 zygotic combinations uniquely characterize S. colombaschense and reveal 5 cytoforms: ‘A’ in the Danube watershed, ‘B’ in Italy’s Adige River, ‘C’ in the Aliakmonas River of Greece, ‘D’ in the Aoös drainage in Greece, and ‘E’ in the Belá River of Slovakia. ‘C’ and ‘D’ are reproductively isolated from one another, and ‘B’ is considered a cytotype of ‘A,’ the probable name bearer of colombaschense. The species status of ‘E’ cannot be determined without additional collections. Three derived polytene sequences, based on outgroup comparisons, place S. colombaschense in a clade of species composed of the S. jenningsi, S. malyschevi, and S. reptans species groups. Only cytoforms ‘A’ and ‘B’ are pests. Within the Simuliidae, pest status is reached through one of two principal pathways, both of which promote the production of large populations of blood-seeking flies: (1) colonization of the world’s largest rivers (habitat specialization) or (2) colonization of multiple habitat types (habitat generalization). Evolutionary acquisition of the ability to colonize large rivers by an ancestor of the S. jenningsi-malyschevi-reptans clade set the scene for the pest status of S. colombaschense and other big-river members of the clade. In an ironic twist, the macrogenome of S. colombaschense reveals that the name associated with history’s worst simuliid pest represents a complex of species, two or more of which are nonpests potentially vulnerable to loss of their limited habitat.

Introduction relationships, and gain insights into the drivers of its historic pest status and the pest status of black flies more generally. By identifying the drivers of pest status, we provide a basis for proactive pest management of black flies, focused on predicting and preventing problems.

Ethics Statement
All collections of larvae and pupae were made on public land with access from public roads. No specific permissions were required to access sites or collect material, and the collections did not involve endangered or protected species.

Collection and Preparation of Material
Larvae of S. colombaschense were collected by hand from 11 sites in 6 countries, mainly during the spring (April-June) but also in the fall (October) ( Table 1, Fig 1), probably representing two generations. They were fixed in modified Carnoy's solution (1:3 acetic ethanol) and held at -4°C until processing.
The posterior half of each larva was severed, opened ventrally with fine needles, and Feulgen stained [4,22], with treatment in cold 5N hydrochloric acid [23]. Larval silk glands with their stained polytene chromosomes, plus one gonad, were transferred to a drop of 50% acetic acid and flattened under a coverslip by applying thumb pressure. Gender was determined by gonadal shape (elongated in females, spherical in males) and confirmed cytologically by the presence of sporadically distributed (female) or clustered (male) meiotic figures.

Chromosomal Mapping and Analyses
Selected high-quality chromosome preparations from larvae collected in the Inn River, Germany, unless otherwise stated, were photographed under oil immersion on an Olympus BX40 compound microscope. Photographic negatives were scanned with a Nikon Coolscan V and imported into Adobe 1 PhotoShop 1 Elements 8 to construct chromosomal maps. Larval carcasses and photographic negatives of chromosomes were deposited in the Clemson University Arthropod Collection, Clemson, SC.
Sections 1-100 of the 3 chromosomes (I, II, III) were numbered on the short (S) and long (L) arms of our maps, according to the Simulium subgeneric standard for the IS, IL, IIL, and IIIS arms [24] and the IIS and IIIL arms [25,26]. Chromosomal landmarks were labeled in accord with those of the subgeneric standard [24]. Banding patterns of stained polytenes of S. colombaschense were compared with those of the subgeneric standard sequence. Inversions fixed across all populations (i.e., interspecific inversions) are italicized in the text and underlined on the maps, whereas floating (i.e., polymorphic) inversions, and any inversions linked to sex, are in Roman type. We used the same numbers for the same inversions found in previously studied, related species; thus, inversions IL-1, IL-2, IIS-1, IIL-1, IIL-2, IIIS-1, IIIL-1, and IIIL-3 of S. colombaschense are identical to those identified previously in related species [21]. We used new, consecutive numbers for novel inversions, and we named each heteroband (hb) by the section in which it occurs (e.g., 6hb).
Autosomal polymorphisms with frequencies of 50% or greater in collections of more than 25 larvae were tested for Hardy-Weinberg equilibrium. All other autosomal polymorphisms were too infrequent for meaningful analysis, occurring in fewer than 10 homologues (typically fewer than 5) per inversion per site.
All rearrangements are indicated by brackets or arrows on the photographic maps, dashed if linked to the X chromosome, dotted if linked to the Y, and solid if fixed or autosomally polymorphic. We identified the sex chromosomes of S. colombaschense by rearrangements (e.g., inversions, heterobands), linked to gender, on one of the three chromosomes (I, II, or III) [27][28][29][30]; these rearrangements were expressed homozygously in females (XX) and heterozygously in males (XY). If no rearrangements were linked to the X or Y, we considered the sex chromosomes to be cytologically (i.e., microscopically) undifferentiated (X 0 Y 0 ). We prepared idiograms to graphically summarize the sex chromosomes and all other diagnostic rearrangements of S. colombaschense.

Terminology
We use the term 'cytoform' as a general descriptor for chromosomally distinct entities that can be recognized at an individual or a population level. The term carries no implications as to whether the entities are interbreeding (cytotypes) or reproductively isolated (cytospecies). 'Pest' is defined here as a species that in some part of its range has caused economic losses or has been a target for management.

Results
The chromosomes of 163 larvae (96 females, 67 males) of 167 that we prepared, including 25 parasitized with unidentified mermithid nematodes, were analyzed completely, band by band, against the Simulium subgeneric standard. The primary nucleolar organizer was in the subgeneric standard position in IIIL. A chromocenter and B (supernumerary) chromosomes were lacking. The total number of rearrangements in the polytene complement, relative to the subgeneric standard, included 59 inversions (12 fixed, 30 floating, and 17 sex-linked; Tables 2 and  3), 10 elaborated bands (e.g., heterobands), 1 secondary nucleolar organizer, 1 transposed nucleolar organizer, and 1 centromeric dimorphism. More than one-quarter of all inversions were sex-linked, all in the distal half of IS, typically atop interspecific inversion IS-2.

Fixed Sequences
The banding pattern of S. colombaschense differed from the Simulium subgeneric standard by 12 fixed inversions: IS-2, IL-1, IL-2, IL-6, IL-7, IIS-1, IIL-1, IIL-2, IIIS-1, IIIL-2, IIIL-3, and IIIL-5. We regard IIIS-1 as essentially fixed although it was absent in 1 homologue of a female larva from Slovakia (Site 9). Fixed inversions in IS of S. colombaschense were nonoverlapping. IS-2 typically was overlapped by one or more sex-linked inversions (Figs 2-4), but by itself moved the '2 blocks'    marker slightly more proximal to the centromere. IL was removed from the standard by 4 fixed inversions, although by 5 inversions in the Danube Basin and Italy's Adige River where IL-5 was fixed or nearly so. These 5 inversions divided the arm into 10 fragments (Fig 5). The most parsimonious reassemblage of fragments to produce the subgeneric standard, one inversion step at a time, is shown below, where the order of fragments is represented by the letters 'a' through 'r,' corresponding with the lettering in Fig 5, and where brackets represent the inversion in each sequence. The order of inversions is constrained by IL-1 and IL-2, which also are found in all or some studied relatives, respectively, and, therefore, must have been the first inversions to have occurred in IL during the evolutionary derivation of S. colombaschense from an ancestor, and by IL-5, which because it is not fixed in all populations, must have been the last inversion in the series.  (Fig 6). IIL differed from the standard sequence by two overlapping basal inversions, IIL-1 and IIL-2 (Fig 7). IIIS carried fixed inversion IIIS-1 (Fig 8), and IIIL had 3 nonoverlapping inversions (Fig 9).
The 12 fixed inversions, relative to the Simulium subgeneric standard, provided the fundamental sequence for recognizing additional diversity within S. colombaschense (Fig 10). We recognize 5 cytoforms ('A', 'B', 'C', 'D', and 'E') among the 11 studied populations. Each cytoform was defined based primarily on the sex chromosomes, supported by autosomal features and a limited number of fixed inversions.

Cytoform 'A'
Populations in the Danube River and its tributaries (Sites 1-3, 8), including the Mureş River (Site 9), which joins the Danube via the Tisa River, all within about 75-500 km of one another, were chromosomally cohesive across space and time (April-October). Cytoform 'A' was characterized by fixation of IL-5, an X chromosome typically based on IS-1, and a Y that carried IS-7. Although males were lacking in our samples from Romania's Cerna and Mureş Rivers, we assigned larvae from these sites to cytoform 'A', based on fixation of IL-5, a probable X chromosome based on IS-1, and the presence of autosomal polymorphisms IIL-13 and IIIC d , which were shared with populations in the Danube River.
Autosomal polymorphisms. Cytoform 'A' had 17 autosomal floating inversions and the highest levels of autosomal heterozygosity among the 5 cytoforms (Table 2). Inversions IIL-14 and IIIL-4 were in Hardy-Weinberg equilibrium for the two seasonal samples from Germany ( Table 2). Three inversions varied clinally from the upper Danube to the Iron Gate: IIL-13 increased from a frequency of 0.02-0.04 in the upper Danube to 0.94 in Romania, whereas IIL-14 and IIIL-4 decreased from 0.50 and higher to 0.00. IIL-14 was associated with a band expression (eb) at the junction of sections 64 and 70 (Fig 7); whether the band was an enhancement (hb) of a fine band or a novel, intercalated band (i) could not be determined. Autosomal heterobands included the pale-staining 31hb in all samples (Fig 5), with frequencies as high as 0.57 per site. Heteroband 71hb in IIL (Fig 7) and a telomeric heteroband in IIIL occurred heterozygously in a female (28 April) and male (24 October), respectively, from Germany (Site 3). The centromere band of chromosome III was heterozygously diffuse (IIIC d , Fig 8) in German larvae (2 females, 1 male) on both sampling dates and in 1 female larva from Romania's Cerna River (Site 8). A male larva from Slovakia (Site 9) had a secondary nucleolar organizer in section 28 (Fig 5), and 1 female larva from the same site had the primary nucleolar organizer transposed, without an inversion, to section 91 on the IIIL-4 homologue (Fig 9).

Cytoform 'B'
This cytoform, from the Adige River in Italy (Site 7), was similar to cytoform 'A' in its basic sequence and X chromosome, which minimally carried IS-1, 3. It was characterized by a Y chromosome with no sex-linked inversions ( Table 3, Fig 2).
Autosomal polymorphisms. IL-5 was nearly fixed (frequency = 0.95). Of the 4 additional autosomal inversions, 3 were shared with cytoform 'A' ( Table 2). Heterobands 31hb and 64hb each occurred (in different larvae) with a frequency of 0.09. The centromere band of chromosome III was heterozygously diffuse (C d , Fig 8) in 1 male larva; this polymorphism was shared with cytoform 'A'.

Cytoform 'C'
Site 4 on the Aliakmonas River in Greece was pure for cytoform 'C,' which was characterized by a high frequency (0.43) of the linked autosomal inversions IS-1 and IS-4 and sex determination based on IIIL. The only triploid larva in our study, a female of 'C', was heterozygous for IS-1, 4, with only 1 of the 3 homologues carrying the inverted sequence.
Autosomal polymorphisms. Three autosomal inversions were found, of which the two most common (IS-1 and IS-4) were linked to one another (Table 2). Heterobands were absent in IS, but 63hb occurred in IIL of 3 larvae (Fig 7).
Sex chromosomes. The typical Y chromosome (Y 3 ) of cytoform 'D' had 4 linked inversions: a complex of 3 overlapping inversions (IS-12, 13, 14) atop IS-2 and the subterminal IS-11 (Fig 4). One larva lacked these inversions; its Y, therefore, was identical to X 0 . A second larva lacked IS-11, but had IS-12, 13, 14 (Y 2 ). Several derivations of the 3 overlapping inversions from the IS-2 sequence are possible; the 3 inversions, therefore, are not individually identified on our maps. Instead, precise breakpoints of all 4 Y-linked inversions (IS-11, 12, 13, 14)  Fig 2, however, are identified on the X-linked IS-1 sequence (although IS-1 is not present in cytoform 'D') and, therefore, the derivation of IS-11, 12, 13, 14 in Fig 2 is from a IS-1 condition, with IS-1 identified on our map by the g|m and h|n breakpoints. The complex Y paired most often (79% of males) with a simple X chromosome (X 0 ; i.e., with no inversions on the fixed IS-2 sequence). Six additional X chromosomes were found (Figs 2 and 3c): X 4 (IS-5), X 7 (IS-9), X 8 (IS-10), X 9 (IS-15), X 10 (IS-16), and X 11 (IS-9, 16). No heterobands were expressed on the X, but Y 2 and Y 3 each consistently had an extra band (i) (Fig 4) intercalated at the k|l junction (Fig 2).

Cytoform 'E'
A single female larva, tentatively recognized as a distinct cytoform, 'E', was collected from the Belá River of Slovakia (Site 11). It was homozygous for IS-1 (tentatively considered X linked), heteroband 31hb, and the unique IIIL-11 inversion (Table 2, Fig 9). It lacked IL-5 and heterobands in IS, setting it apart from cytoforms 'A' and 'B'.

Evolutionary Relationships
One or both breakpoints of 11 inversions shared by 2 or more cytoforms of S. colombaschense could be evaluated in the outgroups: IS-1, IS-2, IS-3, IS-4, IS-6, IL-5, IL-6, IL-7, IIL-12, IIL-14, and IIIL-4 (Fig 11). Ten of these inversion sequences were absent in the outgroups. IS-2, however, was identical to or differed at most by one thin band from the sequence in S. vittatum; we, therefore conservatively treated IS-2 as the same sequence as in S. vittatum. The scrambled distal portion of IS of S. erythrocephalum did not allow us to evaluate the presence or absence of IS-2.
The 5 cytoforms of S. colombaschense formed a monophyletic group united by synapomorphic fixed inversions IL-6 and IL-7. All cytoforms, except 'D', minimally shared IS-1, which was X linked in 'A', 'B', and possibly 'E'. Cytoforms 'A' and 'B' were sister taxa, based on the presence of IL-5, IIL-12, and IIL-14, and by further differentiation of the X chromosome.

Cryptic Taxa
Our uppermost collection of S. colombaschense in the Danube system is from the Inn River at the confluence with the Danube. From here downriver to the Balkan states, including the area of historical abundance at the Iron Gate, the Danube is colonized by S. colombaschense [32]. Our repeated efforts to find S. colombaschense at the Iron Gate have been unsuccessful. However, larvae of cytoform 'A' from Romania's Cerna River, a Danube tributary within about 15 km of the Iron Gate, suggest that cytoform 'A' represents true S. colombaschense of historical notoriety. Similarly, the seven names synonymized with S. colombaschense [19] on the basis of morphological evidence, and applying to populations along the Danube, are supported by chromosomal evidence. Cytoform 'B,' the Adige River population, might represent a distinct species characterized chiefly by its undifferentiated Y chromosome (Y 0 ). However, the suite of chromosomal features shared with 'A'-fixed inversions, X chromosomes, and common autosomal polymorphismssupports a conservative view that 'A' and 'B' are a single species with geographic polymorphism of the Y chromosome. Sex-chromosome polymorphism is common within species of the Simuliidae [5,29]. We, therefore, regard 'B' as a cytotype of 'A'. Our sample of 'B' from the Adige River was taken about 380 km from our nearest collection site in the Danube Basin. We also have ethanol samples, not amenable for chromosomal analysis, from Fiume Rienza (46°4 8'55"N 11°44'12"E) in the Adige catchment and from the Isar River (47°30'22"N 11°17'16"E) in the Danube catchment, only 85 km apart, separated by the northern and central eastern Alps. However, high mountains rising to more than 3000 m between the Adige and Danube populations suggest the possibility of some degree of isolation of 'B'.
Populations of cytoforms 'C' and 'D' in Greece, although as few as 50 km apart, are pure and distinct from one another. Among the 78 analyzed larvae, we found no evidence of interbreeding. IS-1 + IS-4 is common in 'C' but absent in 'D', IIIL-4 is Y linked in 'C' but fixed in 'D', and the sex chromosomes are nonhomologous (I versus III), a classic indicator of separate species [33]. Different species rarely share the same sex-chromosome system, although rearrangements that are sex linked in one species are frequently autosomal or fixed in another species [28]. Sex-chromosome differentiation has been postulated as a driver of speciation [30]. The Aoös and Aliakmonas drainages are separated by mountains that rise to 2400 m, although whether the mountains provide a barrier to dispersal is not known. Given the physical proximity of 'C' and 'D', without evidence of interbreeding, we suggest that they are full species, although their status vis à vis the other cytoforms is not known. Our collections of 'C' and 'D' are about 500 km south of the nearest sampling site for cytoform 'A'. Neither of the two cytoforms in Greece is considered a pest. On the contrary, S. colombaschense had not been recorded in Greece before 2012 [34].
Establishment of a separate cytoform ('E'), based on a single specimen from Slovakia's Belá River, might seem unusual. The specimen, however, is chromosomally distinct and geographically remote (ca. 200 km) from the nearest Danube population. The Belá River is a tributary of the Váh River, which flows into the Danube. The Váh River at one time might have been suitable for colonization by S. colombaschense, but its riverbed and discharge patterns have been modified by a series of hydropower plants, and before 1990 the Váh was heavily polluted, especially by a pulp mill in its upper reaches. Notwithstanding the possible historical connection of populations of S. colombaschense in the Danube and the Belá, the recognition of cytoform 'E' carries the prediction that IL-5 is absent and IIIL-11 is fixed. The situation is reminiscent of cytoform 'CKL' of the Nearctic S. tuberosum complex, which initially was recognized as a possible sibling species on the basis of two larvae [35], and later was shown to be a valid species, S. perissum Dyar & Shannon, when more material came to hand [4,36].
The taxonomic status of two putative species structurally similar to S. colombaschense-S. liriense Rivosecchi and S. voilense Sherban-is relevant to our investigation. Simulium liriense was described from the Liri River of Italy [37] about 480 km south of our sampling site in the Adige River and about 660 km across the Adriatic Sea from our collection site in the Aoös River. Thus, the name might be associated with cytoform 'B', 'C', or 'D', or with a chromosomally distinct entity. The species has not been collected since 1967 from either of the two rivers, the Liri or the Garigliano, where its immature stages originally were found [14], and our two attempts to collect it were futile. The Sacco River, Italy's most polluted river, flows into the Liri, further degrading the habitat and setting up the possibility that S. liriense is extinct [38]; however, additional prospecting is warranted before making a final pronouncement on its status.
Simulium voilense was described from the Doftana River in Romania [39]. The primary diagnostic character used by most authors, in relation to S. colombaschense, is the pupal gill of 10, rather than 10-16, filaments [40]. In a June 1995 sample that we collected from the Danube below the mouth of the Inn River, 57% of 97 pupae had 10 filaments per gill, 8% had 11, 30% had 12, and 5% had 13. This filament distribution is similar to that in the Austrian Danube in April 2011 [40]. Our chromosomal sample of 28 April 2011 from the Inn River (cytoform 'A') had 3 larvae with dark gill histoblasts of which 1 larva had 11 filaments and 2 larvae had 12 filaments; the 24 October sample had 6 larvae with dark gill histoblasts of which 4 had 10 filaments and 2 had 12 filaments. We, therefore, suspect that reports of S. voilense in Slovakia [41] and Italy [42,43], which are based solely on the number of gill filaments, refer to true S. colombaschense. An electrophoretic analysis could not differentiate Italian pupae with 10 filaments from those with 12-14 filaments [44]. The variable filament number in chromosomally cohesive populations demonstrates that the number of gill filaments is not a diagnostic character for S. voilense even though in our material (n = 104), specimens with 10 filaments (58.7%) and  12 filaments (32.7%) represent two clusters, with only 3.8% having 11 filaments and 4.8% having 13 filaments. An even number of gill filaments is disproportionately represented across the Simuliidae [4]. The validity of S. voilense as a distinct species remains in question, and will require comparative material from its type locality in Romania.

Evolutionary Relationships
Simulium colombaschense is a member of the S. jenningsi-malyschevi-reptans clade. This lineage is uniquely defined by inversions IIL-1, IIL-2, IIIS-1, IIIL-2, and perhaps also IL-1 and IIS-2 [21]. IL-2 is shared minimally with S. acrotrichum Rubtsov [21] and the S. reptans complex (unpublished). The occurrence of one female Slovakian larva heterozygous for IIIS-1 is anomalous; the inversion is fixed in all previously studied members of the S. jenningsi-malyschevireptans clade. Either the Slovakian larva carried a mimic inversion of IIIS-1, with imperceptibly different breakpoints, or the standard sequence is a rare, ancestral polymorphism in some clade members.
IS-2, although unique among chromosomally examined species of the subgenus Simulium, is similar, if not equivalent, to the standard sequence in S. vittatum. The IS-2 sequence, or its mimic, is also present in Neotropical taxa, such as the S. (Psilopelmia) perflavum group [45], S. (Psaroniocompsa) quadrifidum group [46], and S. (Psaronicompsa) inaequale group (unpublished). The absence of IS-2 in all other examined members of the S. jenningsi-malyschevireptans clade, and more inclusively in all studied species of the subgenus Simulium, suggests that IS-2 in S. colombaschense is a mimic inversion rather than uniquely shared with S. vittatum and other nonmembers of the subgenus Simulium.
IIIL-5 is shared with the S. reptans complex (unpublished) and possibly also with S. acrotrichum; the breakpoints of IIIL-5 are nearly identical to those at the o|g and p|j breaks within the IIIL inversion complex of S. acrotrichum (fig 8 of Adler & Huang [21]). IIIL-5, however, is lacking in the S. arcticum complex, S. defoliarti, and S. murmanum (unpublished). Thus, among chromosomally known species, the S. reptans complex and S. acrotrichum are in a trichotomy with S. colombaschense. Simulium acrotrichum and S. colombaschense are allopatric, but the S. reptans complex is sympatric with S. colombaschense over the entire range of the latter.
The IS-1 inversion provides a classic example of the different fates an inversion can take in the differentiation of evolutionary lines from an ancestral population [24,26]. The inversion is X-linked in 'A', 'B', and possibly 'E'; autosomally polymorphic in 'C'; and absent (lost?) in 'D'. IIIL-4 also appears in different roles: Y-linked in 'C'; fixed in 'D'; autosomally polymorphic in 'A' and 'B'; and possibly absent in 'E'.

Paradox of Pests and Conservation
The macrogenome of S. colombaschense reveals an irony: The name S. colombaschense, long associated with history's most notorious simuliid pest, includes two or more nonpest species (e.g., 'C' and 'D') that could be vulnerable to loss (extinction?) if their limited breeding habitats are corrupted. The reality of this possibility is demonstrated by the apparent loss of S. liriense from much, if not all, of its original habitat. If S. liriense is a unique species or cytoform and is now lost, it is one of a few examples [49] of a black fly driven to extinction by human agency. Even the superabundant S. colombaschense was eradicated from its original pest-producing habitat in the Iron Gate by habitat alteration, viz., impoundment [9].

Origins of Pest Status
In the Simuliidae, pest status typically is reached by either of two pathways: (1) colonization of large continental rivers more than 100 m wide (specialization) or (2) colonization of a wide variety of smaller streams and rivers (generalization). The essential determinant of pest status in each case is the production of population levels sufficient to exceed an economic threshold. Either pathway can be enhanced by the construction of impoundments with epilimnetic release [3,50,51], which increases food quality and abundance [52].
Of the roughly 40 most significant simuliid pests of humans and domesticated animals [6], about one-third are breeding specialists of the world's large rivers, including seven species in the S. jenningsi-malyschevi-reptans clade. At least 14 (18%) of the 77 nominal species in this clade colonize some of the world's largest rivers, including the majority of species (S. colombaschense, S. kurense Rubtsov & Djafarov, S. luggeri Nicholson & Mickel, S. reptans (L.), and S. vampirum Adler, Currie & Wood) responsible for simuliotoxicosis [7,20]. Among the five cytoforms of S. colombaschense, Cytoform 'A,' the pest of historical infamy, inhabits the largest rivers (e.g., Danube), whereas three of the cytoforms (e.g., 'C', 'D', and 'E') in the smaller rivers are not pests. Of the nine big-river species in the clade, which have not caused simuliotoxicosis, at least three (S. cholodkovskii, S. jenningsi Malloch, and S. penobscotense Snoddy & Bauer) are major, but nonlethal, annoyance pests of humans and domesticated animals [4,20]. Notable among them is S. jenningsi, which colonizes the biggest rivers of eastern North America, and has been the target of one of the world's largest pest-management programs for simuliids [4]. Evolutionary acquisition of the ability to colonize the world's great rivers, such as the Danube, thus represents a key step toward pest status: larger rivers produce larger simuliid populations.
Another one-third of the most significant pests are habitat generalists, colonizing a majority of small streams and mid-sized rivers in an area. These pests include species such as P. mixtum Syme & Davies, S. erythrocephalum (De Geer), S. ornatum Meigen, S. pertinax Kollar, S. venustum Say, and S. vittatum Zetterstedt [4,53,54]. Some of these habitat generalists sporadically colonize the largest rivers in an area and also can become major pests via the big-river pathway [50,[55][56][57].
Available evidence suggests that simuliids evolved in small streams of mountainous terrain [5], the predominant breeding habitat of extant members of the entire family. Colonization of rivers wider than 10 m is characteristic of the majority of species in the S. jenningsi-malyschevireptans clade. Outgroup comparisons indicate that the ability to colonize rivers, including Earth's largest, arose in an ancestor of the S. jenningsi-malyschevi-reptans clade. Additional, independent origins of large-river colonization appear infrequently and sporadically in the phylogeny of the Simuliidae, and include other pests of historical prominence, such as  [53]. Metacnephia lyra (Lundström), a spectacularly abundant species in large, unregulated rivers of northern Fennoscandia [58], illustrates the absence of pest problems when feeding habits do not conflict with human interests; M. lyra feeds on wild birds [59]. The influence of M. lyra on the lotic ecosystem and the native wildlife of an area, however, is considerable [60]. The effects on riverine fauna and local wildlife by the historically important big-river pests must have been similarly profound, particularly before introduction of domesticated animals provided alternative hosts.
Big-river colonization is associated with a set of structural, developmental, and behavioral characters. Structural characters include short, robust labral fans and antennae [4,61], and a boot-shaped cocoon that protects the gill filaments from swift currents and abrasive sediment. The larval bodies of nearly all members of the S. malyschevi and S. reptans groups gradually expand posteriorly, which permits close packing of individuals on substrates [4], further enhancing the ability to achieve large populations. Large-river species without gradually expanded larval bodies, such as S. jenningsi and S. penobscotense, exploit the increased attachment areas afforded by the three-dimensional architecture of beds aquatic plant, such as Potamogeton [62,63]. Colonization of large rivers provides access to rapid delivery of abundant, small food particles that favor dense populations [64] and decrease development time [58], facilitating production of multiple generations. Multivoltinism, in turn, promotes pest status by building up population levels over a season. The pests of the S. jenningsi-malyschevi-reptans clade are typically multivoltine, except S. vampirum, which has one generation annually with three cohorts, giving the appearance of multivoltinism [65]. The ability to inhabit the swiftest areas of big rivers confers the additional benefit of protection from predators [58].
Behavioral adaptations for large-river colonization, such as those related to dispersal and oviposition, are insufficiently understood. Some evidence suggests that big-river species travel far (> 50 km) for a bloodmeal [3,62,66]. Dispersal distances of more than 300 km from the river channel have been reported for host-seeking females of S. colombaschense [9]. These claims, however, should be viewed cautiously; at the time, S. colombaschense was not known from areas of the Danube beyond the Iron Gate. If the dispersal distances are reliable, the situation is reminiscent of some savanna members of the African S. damnosum complex that disperse more than 500 km [67].
We suggest that site fidelity is related positively to habitat specialization and, therefore, should differ between large-river specialists and stream-habitat generalists. As the size of a watercourse increases, the prevalence of watercourses of the corresponding size decreases; thus, the largest rivers are numerically scarcest. Selection should favor females that locate appropriate sites for preimaginal development. Perhaps the greatest assurance of finding appropriate habitat would be a return to the natal habitat. The return of females to natal waterways to oviposit would concentrate and build populations in particular habitats, exacerbating pest problems. It also would promote site fidelity and ecological isolation, potentially leading to population differentiation, including speciation [68]. River-specific chromosomal profiles for the cytoforms of S. colombaschense reflect a pattern of dispersal along, rather than across, watercourses for oviposition. The same phenomenon has been reported in the S. pictipes group, which includes specialists of swift rivers with large rock outcrops [33]. Within the S. tuberosum group in North America, the species showing greatest site specificity, S. tuberosum s. s. (cytoform 'AB'), inhabits larger streams and rivers than do other group members [35]. In contradistinction, habitat generalists, such as the pest species P. mixtum and S. venustum, show little, if any, site fidelity [69,70].

Conclusions
The advantages afforded by colonization of the largest rivers conspire to build pest populations that bankrupt the mind: 7 billion pupae of the North American cattle killer, S. vampirum, were estimated in one rocky weir across the North Saskatchewan River [71], and nearly 1 billion adults of S. jenningsi were estimated to emerge per km per day from large rivers in the eastern United States [62]. Populations of S. vampirum were estimated, before management, to be capable of withdrawing 2 liters of blood from every head of cattle in a 75,000-km 2 area of Saskatchewan [72].
An understanding of the evolutionary factors that drive pest status can focus the search for genes responsible for the enabling traits. Such studies should begin with an assessment of biodiversity within the taxon of interest, sorting out populations, or species, hidden in the genome, which do not contribute to the pest problem, or perhaps are in danger of extinction, with consequent loss of ecosystem function [60]. The big-river specialist S. colombaschense provides a telling example of hidden taxonomic diversity and the associated differences in pest status, all masking as a single pest species. With the rapid progress in genomics of the Simuliidae [73], the genes associated with large-river specialization could be identified. Given structural and behavioral characters that facilitate the production of pests, coupled with an understanding of their underlying genetic basis, a new era of proactive pest management can be foreseen in which pest problems are predicted and prevented.