Paralluaudomyia, Clastrier, 2024

Borkent, Art, 2024, The Phylogeny of the Genera of Biting Midges (Diptera: Ceratopogonidae) of the World, Zootaxa 5438 (1), pp. 1-274 : 242-259

Paralluaudomyia —The monotypic African genus Paralluaudomyia was described by Clastrier (1960) from a single female, noting that P. maculata Clastrier was similar to species of Alluaudomyia but distinctive because of an unusually elongate, single radial cell that extended to the apex of the wing. Wirth et al. (1974) considered it a subgenus of Alluaudomyia . Wirth & Grogan (1988) in their masterful treatment of the Ceratopogonini of the World,

treated it once again as a full genus but pointed out that P. maculata shared a complex wing pigmentation pattern with that of the Alluaudomyia xanthocoma species group ( Wirth & Delfinado 1964). This pattern included two large dark spots, one basal and one distal to r-m, noting that the distal spot is “located where the end of the radial cell would be in the species of the xanthocoma group”. Furthermore, study here shows that all Alluaudomyia (and numbers of other genera) have a pair of campaniform sensilla at the apex of R 3; in all Alluaudomyia this is darkly pigmented and in those Alluaudomyia with further discrete pigmentation there is a larger spot extending onto the membrane. In Paralluaudomyia , with its elongate R 3, this spot is also present but only at midlength (at about where it is on the wing of some Alluaudomyia ) and this spot also includes two campaniform sensilla, strong evidence that the elongate radial cell is an elongation from that point. There are further subapical spots on the wing of P. maculata :

along M 2 and one at apex of CuP that are present in at least some species of the xanthocoma species group. The light spots at the apex of M 2 and near the apex M 4 are also present in some other Alluaudomyia . The complex pattern of pigmentation shared between Paralluaudomyia and some Alluaudomyia strongly suggest that Alluaudomyia is paraphyletic in relation to Paralluaudomyia with its strikingly elongate radial cell. Further to this, as noted above under autapomorphies of Alluaudomyia , the female labrum is apically rounded in some species of A. ( Alluaudomyia )

and A. ( Paralluaudomyia ). All other Ceratopogoninae , other than some Parabezziini and Echinoheleini , have a terminal posteriorly directed peg and this is present in many Alluaudomyia . The apically rounded labrum of some

A. ( Alluaudomyia ) and A. ( Paralluaudomyia ) is further evidence that A. ( Paralluaudomyia ) is likely a derived member of Alluaudomyia . As such, this evidence indicates that Paralluaudomyia should be considered a synonym of Alluaudomyia but as a subgenus at the present time, awaiting study of further material (the holotype is in terrible condition, other than the wings and legs) and a phylogenetic study of the species of Alluaudomyia . The type species of Paralluaudomyia is once again named Alluaudomyia maculata ( Clastrier 1960) but with recognition that it is in a separate subgenus, new status.

Sinicohelea —This monotypic Chinese genus, known only as a single male, was simply placed in the subfamily Ceratopogoninae , without any suggestion of where it might more specifically belong ( Wang et al. 2012). The brief description and limited illustrations give few clues. The hind tarsomere 1 was not described and the illustration of the very base of the tarsomere (their Fig. 3 ) does not show any setae; it is therefore uncertain whether there are palisade setae or not (their presence being indicative of the sister group of the Culicoidini ). However, the presence in this male of elongate flagellomeres 10 and 11, with 10 being longer than 11 (described as segments 12 and 13 respectively; flagellomeres 12–13 were missing), indicates a Forcipomyia , where the feature is nearly unique (also present in some Dasyhelea ). The presence of “1–2 striated simple or pectinate scales on costa and radius” is unique within Forcipomyia , present in only some subgenera (e.g. F. ( Schizoforcipomyia Chan & LaRoux, F. ( Typhonomyia )). The strikingly elongate medial lobe extending posteriorly from sternite 9 is also known in Forcipomyia (e.g. F. Herakleohelea Debenham )) but also in some other genera (e.g. some Dasyhelea , some Kolenohelea , some Stilobezzia ). Here, the genus is recognized as a subgenus of Forcipomyia , new status. The single included species is now named Forcipomyia (Sinicohelea) xuanjui (Yu, Wang & Chen, in Wang et al. 2012). New combination.

Sphaerohelea —A monotypic genus represented by two females from Argentina, this genus is placed in the Sphaeromiini . The presence of batonnets on the fifth tarsomeres indicates that it belongs to the lineage defined by character 101. The lack of the synapomorphy grouping nearly all Johannsenomyiini (character 107: each claw with a stout outer tooth; lacking in Austrosphaeromias and the tentatively placed Lanehelea ) and lack of synapomorphies grouping Palpomyiini + Stenoxenini , places it in Sphaeromiini but based on shared plesiomorphies.

Wannohelea —This monotypic Chinese genus is known only from the holotype female ( Yu et al. 2005). There are few clues as to where this genus belongs. Yu et al. (2005) considered the row of strong spines on tarsomere 5 of each leg to indicate that this genus was a member of the Sphaeromiini ( Borkent 2015) . However, as discussed under character 101, these are not batonnets, which are apically blunt and in double rows. Some genera of Palpomyiini do have these spines, including Amerohelea , Phaenobezzia , some Bezzia , some Palpomyia and Pachyhelea suggesting that Wannohelea is a member of the Palpomyiini . The presence of setae on the anterior anepisternum (see character 69) is present in various genera in Hebetulini + Johannsenomyiini + Sphaeromiini + Palpomyiini + Stenoxenini . When present in Hebetula and Sphaeromiini , however, they are situated near the posterior margin of the sclerite, while in Wannohelea they are located in the middle, similar to that in some Johannsenomyiini + Palpomyiini + Stenoxenini . Setae, when situated laterally on the katepisternum, are present in Mackerrasomyia , Mallochohelea , Phaenobezzia , some Palpomyia and all Stenoxenini (these rather fine setae/ spicules). The presence of thick setae on the scutum is present in some Johannsenomyiini and some Palpomyiini + Stenoxenini (character 60). Wannohelea was reported to have no inner or outer claw teeth. All Palpomyiini have inner teeth but these can be quite small and difficult to discern and this feature should be re-examined in Wannohelea . As simply drawn by Yu et al. (2005) the claws on each leg are not fused, a synapomorphy of Palpomyiini + Stenoxenini (see character 106). The absence of abdominal tergal apodemes in this genus, a synapomorphy of Palpomyiini + Stenoxenini (character 133), is not conclusive because recently emerged females do not have these developed ( Borkent & Craig 1994). Overall, I consider this genus to be a member of the Palpomyiini with a distinct possibility that it is a somewhat modified member of Phaenobezzia , with its tarsomere 5 spines, a single radial cell and abundant thoracic spicules/ setae.

Previous Phylogenetic Analyses

This section summarizes previous literature that has presented an evolutionary tree of some sort. Most of these are based on similarities but some may yet provide potential information for future studies of specific groups. In nearly all previous publications describing new genera, authors note character states that are different or similar to other genera, without consideration of the polarity of the features. Some of these are referred to in the discussion of characters above and are not further discussed in this section. Papers that only discuss classification or characterize higher categories (i.e. tribes, subfamilies) are also not included here but some are included below in “Comments on Some Relationships Presented Here and Discussion of Classification of Genera”.

I and, at times, coauthors, have provided morphologically based cladistic phylogenetic interpretation within a genus ( Borkent & Bissett 1990; Borkent & Grogan 1995; Borkent & Craig 1999; Borkent 2001), a limited group of genera ( Grogan & Borkent 1992; Borkent & Craig 1994; Borkent 1996, 1998, 2000b, 2019b; Borkent & Picado 2008, Borkent et al. 2008 b, 2013; Spinelli et al. 2013), or of broader lineages ( Borkent & Craig 2004; Borkent 2000a, 2014, 2019a). These publications are incorporated in the analysis portion of this paper and are not discussed further here. Otherwise, only a few, relatively recent studies are presented in cladistic terms and these are included below.

There have been only a few attempts at molecular phylogenies for Ceratopogonidae . An early publication by Beckenbach & Borkent (2003) was based on a single gene cox2 and included 14 species representing 12 genera and all (at that time) five subfamilies of Ceratopogonidae , along with six representatives of other nematoceran families. Various analytical parameters and weighing schemes produced a few different trees but the one that included only non-synonymous first positions, all second position variation, plus third position transversions and was constructed by neighbor-joining showed some resolution for a few lineages such as Leptoconopinae being the sister to all other Ceratopogonidae (for most of the results) and the relationships between the three genera of Forcipomyiinae (as defined here). However, even these results were not as objective as the paper might suggest. The paper was initially designed to utilize unnamed specimens for molecular analysis and to compare this to morphological evidence. When it became apparent that some results were skewed in relation to the morphologically-based phylogeny, other forms of molecular analysis were applied to reach a level of conformity.

There have been numbers of “phylogenies” presented of various groups of Culicoides but these are actually single (or a few) gene evolutionary reconstructions based on phenetic similarity (= genetic distance) and as noted by Borkent (2018) cannot be considered as phylogenetic. They are single gene evolutionary reconstructions using phenetic (i.e. genetic distance) rather than character based data and are generated by a simplistic algorithm.

Strandberg & Johanson (2016) studied fragments of five genes ( CAD, TPI, AATS, PGD, CO1) for 32 out of 109 extant genera (then with 111 genera), which included all subfamilies and tribes recognized here other than Washingtonheleini , Parabezziini and Neurobezziini . Their Bayesian analysis mirrors some of the conclusions here but with significant differences in numerous details. The meaning of this is obscure, considering that current standards of gene sequence interpretation have by-passed the use of such a limited number of genes. For example, Leptoconopinae was considered paraphyletic and the relationships between the genera of the paraphyletic Ceratopogonini differed from that presented here. The relationships between the genera in Heteromyiini + Hebetulini + Johannsenomyiini + Sphaeromiini + Palpomyiini + Stenoxenini also is markedly different than the phylogeny presented here. They considered both Forcipomyia and Bezzia as paraphyletic and Palpomyia as polyphyletic and this is a distinct possibility as far as the morphological evidence is concerned (as noted above).

Tóthová et al. (2006) studied relationships between 16 species representing 6 genera and 3 subfamilies of Ceratopogonidae with a single gene, 16s rDNA. They concluded that the subfamilies and tribes were valid but these results, provided in an abstract, have not been subsequently published with any details (including the names of the taxa involved).

The morphologically based studies below are organized from those with a general overview of the family to a generally phyletic arrangement (i.e. starting with early lineages).

The first cladistic study of Ceratopogonidae was presented by Remm (1975), including 18 genera on the basis of 21 adult characters with these listed and presented in a matrix. All subfamilies other than Leptoconopinae (considered a separate and unrelated family) and all tribes other than Washingtonheleini , Parabezziini, Neurobezziini, Hebetulini and Stenoxenini as recognized here, had representative genera. Puzzlingly, he analyzed his matrix by examining how many synapomorphies were held in common by constructing another matrix showing the number of synapomorphies present in each pair-wise comparison. He outlined with boxes which had the highest number and recognized that these corresponded to the three subfamilies he recognized: Forcipomyiinae (including also Dasyhelea ), Ceratopogoninae (combining Culicoidini and Ceratopogonini ) and the newly recognized Palpomyiinae (including Heteromyiini , Sphaeromiini , and Palpomyiini ). Remm (1975) lamented the lack of fossil data which he considered to be the major criterion for polarizing features. Instead of using outgroup comparisons to polarize the characters, he described and utilized a hypothetical ancestor to represent the groundplan, although it is uncertain how its features were determined. All of Remm’s (1975) characters are either utilized as synapomorphies above or are considered as “Uncertain Character States”. As a result of his characterization of the hypothetical ancestor, Remm (1975) concluded that his Palpomyiinae (including what is recognized here as Heteromyiini , Sphaeromiini (s. lat.), Palpomyiini , Stenoxenini ) was the earliest lineage in the family and his Ceratopogoninae (here the Culicoidini and Ceratopogonini ) was the sister group of Forcipomyiinae , thereby inverting the entire classification of the family, contrary to all previous and subsequent classifications and subsequent cladistic analysis.

The first cladistic study of Ceratopogonidae that provided outgroup comparisons was published by Borkent et al. (1987) interpreting the basal lineages of the family and concluding that Leptoconops was the sister group of all other Ceratopogonidae and that Austroconops was the sister group to those remaining. Forcipomyiinae , Dasyheleinae (as recognized then) and Ceratopogoninae were presented as an unresolved trichotomy. Borkent (1995) presented 54 characters interpreting the generic relationships of Ceratopogonidae of all basal lineages, including Ceratopogonini (s. lat.) and treated as unresolved the Heteromyiini , Sphaeromiini (s. lat.), Palpomyiini and Stenoxenini .

Szadziewski (1996) provided a cladogram based on 30 synapomorphies showing relationships between Austroconops , Leptoconops, Forcipomyiini, Dasyheleini , Culicoides , Washingtonhelea , Ceratopogon and Heteromyiini + Sphaeromiini (s. lat.) + Palpomyiini + Stenoxenini and placing nine Cretaceous fossil genera. He did not provide outgroup comparisons for the synapomorphies he presented and Borkent (2000a) discussed these results in some detail.

Díaz et al. (2021) presented a cladistic study of genera of Ceratopogonidae based on immatures, 44 of which are known as pupae and 29 as larvae ( Allohelea was not included even though it has been described by Glukhova (1979)). They included the 83 pupal and three adult characters utilized by Borkent (2014) and added 37 characters for larvae. The outgroup was defined as the groundplan state for Culicomorpha (as was done by Borkent (2014)). One of their cladograms, based solely on the larval features, produced relationships which are highly inconsistent with what is otherwise known about generic relationships. For example, Dasyhelea is presented as the sister group to all other Ceratopogonidae , Forcipomyiini as the sister to Leptoconops + Ceratopogoninae , and Ceratopogon being sister to Schizonyxhelea + Lanatomyia + Phaenobezzia at the apex of the cladogram. The reason for such a high discrepancy is at least strongly related to how the character states were scored. As is standard for most cladograms, the characters were merely tabulated and not discussed (a practice criticized by Borkent (2018)). Some of the characters described by Díaz et al. (2021) are included here but most others were not because of serious problems with their interpretation. In Díaz et al. (2021), 11 of the 37 larval characters were sensilla and the homologies of most are not understood in other families of Culicomorpha (and therefore cannot be legitimately polarized) (but see characters 195–197 here). This is also true for three characters of the epipharynx and the two of the hypopharynx (see characters 198, 199 here). Other characters were misrepresented. For example, their character 36 stated as the presence of an anal pseudopod and considered a synapomorphy of Forcipomyiini , is similar in many Forcipomyia and Atrichopogon to the anal hooks in Austroconops and Dasyhelea (discussed here as character 201); their character 37 is also incorrectly scored in this regard. Their consideration of a prognathous head capsule (their character 1) being derived is actually present in most other Culicomorpha (see character 190 here). So too, for example, with their character 2 (head capsule pigmentation well developed), character 4 (thin setae on head capsule), their character 5 (shape of head capsule, with cylindrical considered plesiomorphic and three states in a morphocline: triangular, conical, quadrangular), character 21 (maxillary palpus long) and character 27 (hypostoma membranous considered as plesiomorphic). Others have similar problems with outgroup comparisons and clearly show the need for discussion of characters and description of states in the various families of Culicomorpha and particularly Simuliidae + Thaumaleidae as the sister group of Ceratopogonidae ( Borkent 2012) . One important larval synapomorphy (red haemolymph) of Austroconops + Leptoconops was not included in their analysis and another (their character 33—body not secondarily divided but probably meant as actually present as it is scored 1 for Leptoconops ) was misrepresented as being present only in Leptoconops when it is also present in Austroconops ( Borkent & Craig 2004) . Synapomorphies of Forcipomyiinae described by Fürst von Lieven (1998) (characters 193, 194 here) were not considered. As such, I consider their use and interpretation of larval features as needing revision, including questions of homologies in other Culicomorpha. Their resulting cladogram alone, being so contrary to previous classifications and cladograms, should be a provocation to do so.

Díaz et al. (2021) also combined their larval characters with the pupal and few adult characters described and polarized by Borkent (2014). Their resultant cladogram largely reflects that by Borkent (2014) with the notable exception of Clinohelea being placed as the sister group to Allohelea + Monohelea + Atyphohelea + Parabezzia and therefore in the Ceratopogonini s. lat. These results are not surprising considering that there were more than twice as many pupal characters than larval and that the pupal characters were interpreted using detailed and broad outgroup comparisons by Borkent (2014).

Hribar & Mullen (1991) presented the comparative morphology of larvae of 12 genera, in some detail, representing each of the subfamilies and all the tribes recognized at that time. They provided illustrations of various structures for each genus and made broad comparisons within the context of the classification. Their paper will be a continued resource for future work interpreting character states of the larvae, especially as outgroup homologies become better understood.

Downes (1977) presented, as part of an abstract, a figure of a phylogeny of the subfamilies and tribes of Ceratopogonidae that somewhat mirrors the one presented here. He considered Ceratopogonini to be monophyletic while here it is paraphyletic. Downes (1977) did not present any morphological evidence for the phylogeny but Downes (1978) discussed numbers of features of the predaceous taxa and presented summary statements and a table summarizing his conclusions. The characters he viewed as significant are discussed about as synapomorphies or as “Uncertain Character States”.

The relationships between the species groups and subgenera of Forcipomyia were interpreted by Chan & LeRoux (1971b) and included a discussion of the relationship between Forcipomyia , Atrichopogon (stated as being “plesiomorphic and the genus a possible living prototype of all Ceratopogonidae ”), and the other groups of Ceratopogonidae . They provided a summary of the subfamilies of Ceratopogonidae recognized at the time, including Dasyheleinae . Forcipomyia and Atrichopogon were noted as the sole members of Forcipomyiinae . Their phylogenetic conclusions were based on considering the male genitalia of Atrichopogon to be “a possible living prototype of all Ceratopogonidae ”. The concluding evolutionary tree, claiming to be the result of Hennigian analysis, was based on divergence of the parameres (as claspettes) and shows Forcipomyia arising six times from within Atrichopogon . Further to this, the Atrichopogon “prototype” was also the separate origin for each of Leptoconopinae (then including only Leptoconops ), Dasyheleinae (here = Dasyheleini ), Culicoidini + Ceratopogonini and of the remaining Ceratopogoninae . This remarkable conclusion is by far the most inconsistent with every other attempt at classification and/or phylogenetic perspective. Confusingly, Chan & LeRoux (1971b) state in their text that Dasyhelea and F. ( Lasiohelea ) (as Lasiohelea ) are probable sister groups, contrary to their phylogenetic tree. Their method of determining polarity lacked outgroup comparisons in other Culicomorpha. A discussion of some of the characters of various subgenera of Forcipomyia included other features with some of these being pupal and larval structures. Much of their summary was based on the magnificent work by Saunders (1924, 1925, 1957, 1959, 1964). In his 1957 paper Saunders described a number of species and four new subgenera and presented a chart (pg. 703) combining adult, pupal and larval characters of each recognized subgenus (at that time), thereby laying the groundwork for the classification for subsequent workers (see especially Debenham 1983, 1987a, b, c, d). Debenham (1987d) provided a detailed critique of Chan & LeRoux (1971b) and her own interpretation of the characters. In particular, she noted the importance of the shift of male pupal genital processes from a ventral to a dorsal position, also noted by Chan & LeRoux (1971b), grouping a number of subgenera of Forcipomyia , and she produced an evolutionary tree reflecting this and presumably the other features she discusses.

Forcipomyiini are particularly rich in character states in each life stage and the group is calling for a cladistic analysis and further study. It would help to interpret the remarkable diversity of feeding habits of the adults and numbers of other features (see for example the amazing website of Andy Murray—https://www.chaosofdelight. org/forcipomyia).

Boorman & Lane (1979) discuss their new monotypic genus Neoculicoides Boorman & Lane (now a synonym of Culicoides , with it single species C. taylori Boorman & Lane considered a miscellaneous member of the genus) in the light of relationships to Paradasyhelea , Austroconops using features which were a combination of various similarities (e.g. small size, simple legs) or strongly homoplastic features most of which are discussed above in “Uncertain Character States”.

Szadziewski et al. (2019a) described the new subgenus Culicoides (Groganomyia) to include one extant species, Culicoides cameroni and four Cretaceous species from 99 million year old Burmese amber. They proposed that the subgenus is the sister group of remaining Culicoides based of the lack of sensilla coeloconica on at least some or all of flagellomeres 2–13 (these are absent in some other Culicoides , see character 25) and the presence of dorsal and ventral gonocoxal apodemes (see “Uncertain Character States”—Male gonocoxal apodeme”; the apodemes also occur in some other Ceratopogonidae ). Based on this evidence alone, their conclusion that C. (Groganomyia) is the sister of all other Culicoides could equally well indicate the possibility that it belongs elsewhere (i.e. C. (Groganomyia) has none of the synapomorphies present in remaining Culicoides ) and they confusingly also indicate the Culicoidini are paraphyletic. They discuss some further character states which are present in some other genera, which are also discussed here, namely: the female interocular sulcus (“Uncertain Character States”—“Female head capsule vertex with transverse sulcus dorsal to the medial seta”), the anterior lateral cervical sclerites (character 49), macrotrichia on the wing membrane (“Uncertain Character States”—“Presence or absence of wing membrane macrotrichia”) and female genital sternite 9 (character 142). Of these, character 49 (slender and S-shaped anterior cervical sclerite) is here considered a synapomorphy of Culicoides and Paradasyhelea , indicating that C. cameroni belongs to at least this group. However, character 142 (female with a medially discontinuous abdominal sternite 9) is present in nearly all Ceratopogoninae (with some reversals to the fused condition, including in a few other Culicoides ); as such C. cameroni with a continuous sternite may be the sister to all other Ceratopogoninae (but here treated as a reversal). If valid, this character indicates that Culicoides and Culicoidini is indeed paraphyletic (because it would not include C. (Groganomyia). Further to this, based on the evidence of the sternite 9 alone, it is possible that C. (Groganomyia) is basal to the Forcipomyiinae . However, the male of C. cameroni has a pedicel with a narrow basal foramen (character 13), indicating it belongs at least to the Ceratopogoninae . Further to this, the female medial suture of the prementum is simple dorsally (see character 47), a feature of other Culicoides as a reversal and likely an autapomorphy of the genus, indicating that the subgenus does indeed belong in Culicoides . Other features of C. cameroni are included in the matrix with Culicoides for adult features (immatures unknown).

Szadziewski et al. (2019a) included only a few adult synapomorphies in their interpretation of C. (Groganomyia) and did not consider a pupal synapomorphy of the genus proposed by Borkent (2014: character 53). Although the single extant species of C. (Groganomyia), C. cameroni , is yet unknown in the pupal stage, other Culicoides , such as C. floridensis and at least C. reconditus Campbell & Pelham-Clinton of the subgenus C. ( Wirthomyia ) with a continuous sternite 9 are known as pupae ( Glukhova 1989; Linley 1970) and should be examined as evidence that Culicoides is monophyletic.Although Linley (1970) did not describe the pupal synapomorphy feature, it is clear that the species looks like a typical Culicoides (including with one long and the other quite small dorsolateral cephalic sclerite setae); this, however, is not a phylogenetic argument and further investigation is needed. Although unknown as a pupa, Campbell & Pelham-Clinton (1960) report C. cameroni “bred from moss”.

Finally, it may be important to point out an apparent autapomorphy of C. cameroni . The male has separate parameres which are closely approximated at their bases and with a small separate median sclerite anterior to their bases, which lies between the gonocoxal apodemes. The median sclerite, separate from the parameres, appears to be unique, as suggested by Campbell & Pelham-Clinton (1960) for the genus, in the family.

The monotypic genus Washingtonhelea was originally described by Wirth & Grogan (1988) and placed in the Ceratopogonini as “one of the most plesiotypic genera of the tribe” and several features “place it near the tribe Culicoidini ”. Borkent & Grogan (1995) placed the genus together with Culicoides , Paradasyhelea and all remaining Ceratopogoninae ( Ceratopogonini et al.) as an unresolved polytomy. Borkent (1995) placed it as an unresolved lineage, along with Forcipomyiinae and all Ceratopogoninae (including Culicoidini ), based on a male pedicel foramen intermediate in size and a plesiomorphic entire female sternite 9 (his characters 17” and 26, respectively) with the apomorphic conditions being synapomorphies of Ceratopogoninae . Both were inaccurate. The two characters are here presented as characters 13 and 142 respectively. Szadziewski (1996) once again recognized Washingtonhelea as a member of Ceratopogonini but without explanation; his transferring of two fossil species to the genus was subsequently refuted (Borkent 2000). Borkent & Wirth (1997), in their world catalog, included Washingtonhelea as one of three genera of Culicoidini (together with Culicoides and Paradasyhelea ) but did not explain that decision and this was presented as such in a phylogeny by Borkent (2000) as a potentially paraphyletic group (i.e. lacking a synapomorphy). Szadziewski et al. (2019a) correctly state that “We do not agree with Borkent (1995, p. 94) that females of Washingtonhelea have an entire sternite 9. The female paratype of W. frommeri that we examined has a narrowly separated sternite 9, and their submedian halves have finger-like projections directed caudally and are almost touching.” (see also other features he discusses). See character 142 for further discussion of this feature. Reexamination of the male pedicel indicates that it actually has a narrow base (character 13). Combined with other features (characters 31, 35, 39), the current evidence indicates Washingtonhelea is the sister group of all Ceratopogoninae other than the more basal Culicoidini and its phylogenetic position is here recognized as a new tribe.

Szadziewski et al. (2019b:661) suggest that the reduction of the laciniae is a synapomorphy of the Ceratopogonini , including Washingtonhelea . However, it is likely that they meant the entire subfamily Ceratopogoninae . The lacinia is discussed further in “Uncertain Character States”. Although likely true, the feature exhibits significant homoplasy in lower lineages.

Grogan & Wirth (1980) proposed that Macrurohelea is the sister group to Ceratopogon and considering the polytomy that includes these two genera near the base of the Ceratopogoninae , this is a possibility. Grogan & Wirth (1979c) suggested that Notiohelea , another genus in this polytomy, was related to these two genera and Spinelli et al. (2009a) also indicated similarities of the male genitalia to those of Macrurohelea . However, these papers do not provide evidence for these relationships, discussing only similarities of features considered to be inadequate here for phylogenetic interpretation (e.g. antennal sensilla coeloconica, number of spermathecae, longer second radial cell, cordiform tarsomeres 4).

Szadziewski et al. (2022) presented a new monotypic fossil genus * Monogedania previously considered a species of Monohelea within the context of the ‘ Monohelea complex’ as defined by Wirth & Grogan (1988). In this work, a table is presented of 25 characters with seven genera ( Monohelea , Downeshelea , Allohelea , Isthmohelea , Schizohelea , Austrohelea , * Monogedania ) scored with the features as either present or absent. All of these are discussed here as either synapomorphies or as “Uncertain Character States”. However, in their table, some characters are actually merely opposite states, such as their 14 and 15 being “tarsomeres 4 of fore, mid legs cylindrical” and “tarsomeres 4 of fore, mid legs cordiform” respectively. The same is true of their 16 and 17 with tarsomere 4 of the hind leg either elongate or short and their 23–25 with the different forms of the aedeagus.

Wirth & Grogan (1988), as part of their comprehensive and synthesizing description of the genera of Ceratopogonini , reorganized the genus Monohelea (giving full generic status to some previous synonyms) and what they considered to be related genera, including at that time Monohelea , Downeshelea , Allohelea , Isthmohelea , Schizohelea and Austrohelea . The group was christened the “ Monohelea complex”. All these genera share a combination of the following: male with 13 flagellomeres; if wing without distinct pigmentation then without a pit on the third palpal segment; palpus with five segments; wing with two well developed radial cells; fourth tarsomeres cylindrical (other than Austrohelea and therefore not identifying correctly in their key); hind femur at most slightly swollen and without stout ventral spines; flagellomere 1 without sensilla coeloconica; females with equal fore- and midleg claws but an elongate hind leg claw. Their table 1 summarized the differences between these genera. The features present in this complex are discussed here, either as synapomorphies or, for most, as “Uncertain Character States”. This explains why this grouping of genera is not considered to be monophyletic here and with the Monohelea complex restricted to the following five extant genera: Boreohelea (previously a synonym of Allohelea, Borkent & Dominiak (2020)) , Monohelea , Heterohelea , Downeshelea and Allohelea (and the fossil genus Monogedania ).

Spinelli et al. (2018) described a monotypic new genus Yungahelea and interpreted it in a cladistic context with other genera of the Parabezziini , as defined here and including at that time, Leptohelea (here transferred elsewhere). They used Bahiahelea , a genus with strongly modified male genitalia and here included as a genus of Parabezziini , as the sole outgroup for their study of the remaining genera. They scored 11 characters and some of these are utilized here and some are discussed under “Uncertain Character States”. Of their four cladograms, their Fig. 24a , in which Spinellihelea is the the sister of Yungahelea + ( Diaphanobezzia + Parabezzia ), is very similar to that presented here, where Spinellihelea , Yungahelea and Diaphanobezzia + Parabezzia form an unresolved trichotomy.

Grogan & Wirth (1979b) presented a phylogeny of the Heteromyiini (including Neurobezzia at that time) and indicated that a “barely sessile media”, here described as M 1 and M 2 forking at r-m, was present as an apotypic state in Neurobezzia , Neurohelea and Physohelea . Neurobezzia was shown by Borkent (1998) as belonging elsewhere. Neurohelea does have M 1 and M 2 forking at r-m but Physohelea females and some males actually have them forking basal to r-m (with some males forking at r-m). Other Heteromyiini have them forking well basal to r-m. Regardless, forking at r-m is an intermediate state (present in other genera as well) and cannot logically be used as a synapomorphy even within the analysis by Grogan & Wirth (1979b); at best it would have grouped all Heteromyiini . Further to this, their synapomorphy grouping other Heteromyiini as “M broadly sessile” (well basal to r-m) is present in Hebetulini and many genera of Johannsenomyiini , Sphaeromiini , Palpomyiini and Stenoxenini . The variation in the position of M 2 is discussed under character 82.

Grogan & Wirth (1981a) proposed a phylogeny of those genera of Johannsenomyiini with female claws having a stout, external tooth. Their phylogeny differs entirely from that presented here but most of the characters they used to indicate lineages are discussed above, either as synapomorphies but mostly as “Uncertain Character States”. One of the characters they discussed was based on their conclusion that r-m was absent in Niphanohelea , a feature otherwise known only in Leptoconops . However, Borkent (2004) presented a different interpretation of the wing as part of a key, without explanation. The wing of Niphanohelea is markedly derived and Grogan & Wirth (1981a) interpreted this as due to loss of r-m and a newly evolved “adventitious” vein. Here I propose a simpler interpretation in which r-m is present but has moved to a subapical position on the wing, shown as two intermediate and hypothetical evolutionary steps in Figs. 78A–D . As such, r-m is retained and the “adventitious” new vein is actually M 1. Furthermore, the thinner, apical portion of M 1 is of similar thickness to M 2 ( Fig. 78E ), supporting the idea that M 1 is not an “adventitious” vein. Generally, in Ceratopogonidae , M 1 and M 2 are of similar thickness, with both significantly thinner than R 3.

Grogan & Wirth (1980), in redescribing the monotypic genus Pachyhelea , presented a phylogeny showing it to be the sister group of Palpomyia (with the four species groups of Palpomyia also interpreted) (also presented in Grogan & Wirth 1979a). Their evidence was that they “were impressed with its resemblance to members of the tibialis group of Palpomyia ” ( Grogan & Wirth 1975) but that because they could list four autapomorphies for Pachyhelea it must be considered a valid genus. However, this is not evidence of a sister group relationship because it includes the possibility that Pachyhelea is merely an autapomorphic species within Palpomyia (and most likely within the tibialis group). Here Pachyhelea is merely part of an unresolved lineage group of four genera within the possibly paraphyletic Palpomyiini ( Fig. 9 ).

Comments on the Classification of Genera and Some Relationships Presented Here

Like every group of organisms, classifications are refined through time, hopefully based on the increasing utilization of characters of phylogenetic importance (i.e. synapomorphies). Although the phylogenetic arrangement and classification proposed here is based on cladistic analysis, portions of the classification as well as some generic relationships present problems or challenges and require some comment. Some individual genera are discussed in a section “Realigned, Synonymized or Poorly Known Genera” and there is further discussion of some genera in “Previous Phylogenetic Analyses” as well.

Wirth et al. (1974) provided an historical synopsis of the higher classification of Ceratopogonidae . The earliest use of subfamilies and tribes was by Wirth (1952), resolving, in part, previous accounts of adults and immatures. Wirth et al. (1974) introduced further tribal divisions that were largely utilized by all subsequent taxonomic publications and certainly provided the backbone of the classification here.

Wirth (1952) erected the tribe Stilobezziini on the basis of a number of features, combining plesiomorphic states and those discussed here in “Uncertain Character States”. According to his diagnosis, in large measure they were Ceratopogonini with female legs having large and unequal claws. Wirth & Grogan (1988: viii) discuss the treatment of Ceratopogonini by subsequent authors who recognized the two tribes, Ceratopogonini and Stilobezziini . As recognized by Wirth & Grogan (1988), the distinction between the two was based on the presence or absence of sensilla coeloconica on the first flagellomere and whether the female claws were unequal on at least one pair of legs. Their study of relatively recently described southern hemisphere genera as well as a “preliminary cladistic study” (not published), concluded that only one tribe was justified and the Ceratopogonini is in large measure the same as that presented here with some exceptions. I consider the newly recognized Washingtonheleini , Parabezziini, Neurobezziini and Echinoheleini (all included genera previously placed in Ceratopogonini ) to be sufficiently supported cladistically and distinctive enough to warrant their tribal status. The remaining Ceratopogonini remain here as a clearly paraphyletic group, with five genera at the base of the Ceratopogonini and several unresolved polychotomies unresolved. Some of the proposed monophyletic groups are based on only a single synapomorphy, with even some of these being homoplastic and/or as, in part, reversals. Because of this level of uncertainty, it is best to leave the tribe as such until further work is done. Even if the phylogeny depicted here is absolutely accurate, the introduction of many further tribal names would likely be a disservice to others studying the group unless distinctive features can be determined.

Within the Ceratopogonini , I have continued to use the “ Monohelea complex” considering that this has been used in several previous papers (e.g. Clastrier & Delécolle 1990, Santarém et al. 2020, Szadziewski et al. 2022) albeit with a different definition than here ( Fig. 6 ) (see further discussion in “Previous Phylogenetic Analyses”).

The position of Schizohelea , also within Ceratopogonini and previously considered a member of the Monohelea complex, is somewhat tenuous although it is distinct from members of complex as defined here. Here it is placed as the sister group of Austrohelea but this is based on the shared reversion to the plesiomorphic state of character 67, in which the anterior point of the medial suture of the anterior anepisternum terminates anterior or ventral to the anterior spiracle. This reversion is not unique but occurs also in other closely related lineages: Fanthamia , Bothahelea , Ankylohelea , Neohelea and Metacanthohelea . There are some further conflicts of character states. Schizohelea male cerci have the most derived state of character 115, otherwise present only in the sister group of the Monohelea complex. Confusingly, the anapleural cleft is narrow or wide (character 70), depending on the specimen—some males and some females had a narrow anapleural cleft but most have a wide cleft, depending at least in part, whether the thorax was partially swollen or not. Here, I consider the wide anapleural cleft as characteristic and have scored it as such. Finally, much of the base of the M 2 is absent making it difficult to score whether it might have originated basal or proximal to r-m (character 82). In pinned specimens there is a fold line from the pigmented distal portion of M 2 that extends to a level basal to r-m. However, Ryszard Szadziewski (pers. comm.) pointed out that there is a dip in M 1 where M 2 normally would originate if the condition was distal to r-m.

The Ceratopogonini genera Cacaohelea and Parastilobezzia are sister groups with two synapomorphies: the cardo and postgena are separate mimicking the plesiomorphic condition but distinct in being lightly pigmented and with the cardo and stipes surrounded by membrane (character 40) and a reduced ventral plate (character 127). They also have a similar labrum with the unusual arrangement of abutting terminal pegs, pointing somewhat medially. The fusion of palpal segments 2+3 of Parastilobezzia , Leptohelea and Camptopterohelea is unique in the family (character 45) and if a valid synapomorphy, conflicts with the conclusion here.

The spermathecae of species of Notoceratopogon have partially wrinkled or at least irregular surfaces, identified as an autapomorphy above of this genus. The spermathecae of Capehelea have a similar shape and are ridged basally and this may be a synapomorphy of these two genera. At present, I have considered Capehelea as having separate antepronotum (not forming a transverse band) described in character 50, the only character state separating the two genera. However, the single male and female of Capehelea I examined had partially crushed thoraces and were perhaps misinterpreted. If further study indicates that the antepronotum is similar in Notoceratopogon and Capehelea , they would be considered as sister groups.

Recognition of the new tribe Parabezziini here, with seven genera, has some issues. First is that Bahiahelea has, at best, only a minor expression of synapomorphy 77 (see under that character). Its consideration as a member of the tribe is partially based on its sister group Parabezziini new genus but readers should know that both have strongly modified male genitalia and perhaps homologies there have been misinterpreted. Second, character 86, the presence of stout spines on the male fore coxa, as a synapomorphy of Parabezzia and Diaphanobezzia is not present in all Parabezzia and that some Parabezzia have stout setae on their midcoxa, as do Diaphanobezzia . Third, Yungahelea have five or more stout, long costal setae basal to the arculus and this occurs in at least males of Diaphanobezzia . Some Parabezzia have up to three stout, long setae. As such, there is a possibility that Parabezzia is paraphyletic in relation to Diaphanobezzia and Yungahelea . Further to this, if the stout spines on the forecoxa can be lost in some Parabezzia , there is a possibility that Spinellihelea also is a derived form of Parabezzia .

The recognition of the new tribe Neurobezziini , including four genera, Meunierohelea , Atyphohelea , Neurobezzia and Ceratopalpomyia , is based on the male antenna having the plume setae on flagellomere 10 arranged as an incomplete (medially) transverse ring, a reversal to the plesiomorphic state from the intermediate stater of the plume setae on flagellomere 10 arranged obliquely at least in part, a feature defining Forcipomyiinae + Ceratopogoninae . Each of the four genera of Neurobezziini are known from one extant species (some with fossils) or, for Neurobezzia , three species and most from very few specimens.

Echinohelea has some interesting features that challenged their placement here. The adults are peculiar in numbers of characters as is reflected in the number of autapomorphies shown in the cladogram ( Fig. 7 ) as well as described in “Autapomorphies of Extant Genera”. Wirth (1994) pointed out that “the genus is distinguished by a number of distinctive apotypic character states, while within the genus the species are remarkably uniform”. Although he states that there had “been no cladistic or phylogenetic discussion of the taxonomic position of Echinohelea ”, Borkent (1995) had previously discussed two characters indicating the same phylogenetic position as indicated here; one of the synapomorphies (his character 53), however, actually now also includes the more basal Neurobezziini (character 82). On the basis of characters 41 (cardo extending dorsally past posterior tentorial pit and with an extension of cuticle dorsal or dorsolateral to posterior tentorial pit) and 64 (anterior margin with lateral apodemes closely approximated medially and the medial apodeme reduced but present) it is considered the sister group of Heteromyiini + Hebetulini + Johannsenomyiini + Sphaeromiini + Palpomyiini + Stenoxenini . Wirth (1994b) discussed several features which he considered important in placing Echinohelea in the Ceratopogonini and these characters are discussed elsewhere. His character 4 includes a statement that “the mesal side of the tergite distally bears a rounded membranous area bearing the cerci, bordered by an incomplete, heavily sclerotized margin laterally and distally. A similar configuration is present in the male genitalia of a number of genera in the tribe Ceratopogonini , including Bothahelea , Brachypogon , Ceratohelea , Ceratopogon , Fanthamia and Kolenohelea ”. It is uncertain exactly what he was referring too but there are significant modifications of the male genitalia that initially suggested to me that they were homologous to a group of genera here placed more basally: Fanthamia , Afrohelea and Bothahelea . In Echinohelea and these three genera, the epandrium is narrow subapically and expanded posteriorly, with the posterolateral, broad extension with a medial or subapical seta (character 117.1). In just Echinohelea and Afrohelea , the epandrium is modified further, with a posterolateral triangular projection, and 1–3 setae near its base (character 117.2). In all four genera the epandrium also has medially directed flanges that are broadly fused medially (or nearly abutting), forming a short “tube”, and with its posterior margin encompassing the anterior margin of the proctiger (character 119) and the female abdominal sternite 8 much larger than sternites 6–7 (character 135). In Echinohelea , Afrohelea and Bothahelea the lateral margin of segment 9 is a very short band or absent, with the hypandrium expanded posteriorly, abutting or fused to the bases of the gonocoxites (character 113). These derived character states suggest that Echinohelea could be the sister of Afrohelea and together be the sister group of Bothahelea , and all three the sister of Fanthamia . Character 12 (male with female-like antenna) indicating the monophyly of Fanthamia , Afrohelea , Bothahelea , Ankylohelea , African new genus A and Calcarhelea is also shared with Echinohelea but this character is otherwise homoplastic and likely a weak indicator of relationship. Wirth (1965), in naming Afrohelea , suggested it was related to Echinohelea but without noting the reason for this conclusion. Wirth & Grogan (1988), however, noted Afrohelea “Superficially resembling Echinohelea in the shape of the male genitalia…”.

The relationships shown for the Heteromyiini may need revision. Pellucidomyia are strongly modified as adults and pupae and aside from exhibiting a number of reversals ( Fig. 7 ), share some derived features with Heteromyia such as the palisade setae on female hind tarsomere 1 extending <0.8 of its length, unique within Heteromyiini (but susceptible to homoplasy, see character 97) and the shared loss of a bifid midtarsomere 4 (character 100), a feature otherwise grouping some Heteromyiini . In addition, Pellucidomyia , Heteromyia and Tetrabezzia are unique in the family in having extremely elongate hind tarsomeres 4 and 5 and a single elongate hind claw with a basal tooth (as in Fig. 77E ) (very tiny and dorsal in Pellucidomyia ), a unique combination within the family (some others have elongate hind claws). Finally, Pellucidomyia females have abdominal sternite 8 with a few curved, medially directed setae ( Fig. 70D ), similar to those otherwise unique to a group of genera in Johannsenomyiini (character 136). Borkent (2014) indicated that Heteromyiini , including Pellucidomyia , was a paraphyletic group based on pupal synapomorphies but not related to Johannsenomyiini .

There are now eight genera placed in the tribe Sphaeromiini genera but with Sphaerohelea and Chelohelea too poorly understood to place phylogenetically. Alloimyia and Indobezzia (neither seen firsthand and known only as females) are successive sister groups to the remaining four genera: Homohelea , Sphaeromias , Leehelea and Xenohelea are confidently considered to be monophyletic ( Fig. 8 ). Whether they should all be recognized as valid genera is another question. Sphaeromias , Leehelea and Xenohelea are here considered to be monophyletic based on the presence of a slender apodeme present on sternite 8 (character 138), difficult to see and missing in some Sphaeromias . The difference between Homohelea and Sphaeromias is minute and reported for females by de Meillon & Wirth (1991: 48) as Homohelea with 1–2 pairs of batonnets on tarsomere 5 and 3 or more for Sphaeromias . Males of the two genera cannot be recognized confidently. Out of 19 known species, male Homohelea are known only from the male syntype of Homohelea telmatoscopa (present on three slides: wings, 1 crushed scutellum) and H. delanoe (de Meillon) known only from the holotype and not studied here. Ten males of Homohelea insons were recorded but not described by Wirth & Ratanaworabhan (1981b). Borkent (2017) in his treatment of Afrotropical genera, keyed out male Homohelea with one species of Sphaeromias on the basis of the parameres being fused, apically bulbous and pubescent. Other Sphaeromias species were keyed twice. The presence of pubescence is likely a synapomorphy of at least some Homohelea (perhaps all) and only some Sphaeromias . The only potentially significant difference between Sphaeromias and Xenohelea (known only as females and pupa of one species) depends on minor tarsal claw differences, with these being reduced in Xenohelea (a phenomena which occurs in some other genera). Leehelea , with nine Oriental and Australian species, was described as new by Debenham (1974). She noted that males have completely separate parameres while these are fused, apically “expanded and rounded, setose” in Sphaeromias . Female Leehelea were characterized as having “simple claws which are not swollen centrally and are as long as or longer than tarsomere V, the ventrally bearded tarsomere IV, and the setae of the legs which are very fine, pale and erect or semi-erect, and arising from distinct microtubercles, rather than dark and bristle-like with normal bases.” However, all but one of these character states are not distinctive. As noted above, some Sphaeromias have separate parameres (e.g. S. eugenei de Meillon & Wirth , S. pistiae (Ingram & Macfie)) that are indistinguishable from those of Leehelea ( de Meillon & Wirth 1987 a, Ingram & Macfie 1922). Regarding females, in the single Leehelea female I studied, L. punctipes (Macfie) , the claws (the longest of the two fused claws) were as long as their respective tarsomere 5. In some Sphaeromias the claws are only slightly shorter, with, for example, S. longipennis (Loew) foreclaw 0.93 the length of foretarsomere 5 and hind claw 0.92 the length of hind tarsomere 5. This level of difference is equal to the intrageneric variation in many genera and cannot be considered significant. So too, the claws of Leehelea may be somewhat more slender than those of Sphaeromias but hardly significant, especially in light of their slightly greater length. The dense spicules on tarsomere 4 (the “ventrally bearded” of Debenham (1974)) does not distinguish Leehelea from Sphaeromias as these are also present in at least S. discolor (de Meijere) , S. fasciatus and S. longipennis . Indeed, this feature is more broadly distributed in numbers of other taxa: see “Shape of tarsomere 4 of each leg, presence of spicules and sinuate setae’ in “Uncertain Character States”. Finally, the “very fine, pale and erect or semi-erect, and arising from distinct microtubercles” of Leehelea (primarily on the femora and tibiae) look somewhat distinctive but, again, only slightly so, with the socketed base of each seta slightly strengthened by additional cuticle. The setae themselves are more slender and elongate than in Sphaeromias . However, this is the only possible distinctive feature, and only of females, so that it appears highly likely that the genus will be considered a synonym of Sphaeromias in the future. I am reticent to do so here because I consider it prudent to await a revision of all four genera, Sphaeromias , Leehelea , Homohelea and Xenohelea as a group, to determine phylogenetic relationships and make more stable nomenclatorial decisions. One pupal apomorphic feature (character 174, flattened tubercles on abdominal segments 3–7) is unique to Homohelea and Xenohelea and possibly synapomorphic (based on one species from each genus). This feature was discussed by Borkent (2014) as character 51. Finally, it is possible that Indobezzia also could be considered as species of Sphaeromias ; see discussion of this genus under “Realigned, Synonymized or Poorly Understood Genera”.

In the phylogeny presented here, four genera, Lanehelea , Austrosphaeromias , Neosphaeromias and Mackerrasomyia are considered to be a monophyletic group within the Johannsenomyiini . When Spinelli (1997) first described Austrosphaeromias , he placed the new genus in Sphaeromiini (s. lat.). Later, Spinelli et al. (2015) noted the following: “During our study of Patagonian predaceous midges in the genus Palpomyia Meigen ( Spinelli et al. 2009b) , we initially considered the latter two above species [ A. apricans , A. wirthi ] as members of this genus because their uncanny resemblance to species of the distincta group. For example, both species possessed greatly swollen fore femur with numerous ventral spines, and females had short, equal-sized claws and small, ovoid subequal-sized spermathecae ( Grogan & Wirth 1979a). However, males lacked a mesobasal tubercle on their gonocoxites as well as a broadly triangular apex on their aedeagus, and females lacked a pair of soft setose lobes on the anterolateral margins of sternite 8 and a distal hyaline plate on that structure ( Spinelli et al. 2015). In addition, sternite 8 of females of the Palpomyia distincta group features a separate proximal portion with scattered setae, whereas both species in question possessed a closely approximated group of more coarse setae.” Although most of the features noted by Spinelli et al. (2015) are either strongly homoplastic or are not known to be apomorphic conditions, their interpretation that the genus might be more closely related to Palpomyia warrants further study. The only synapomorphy known for Lanehelea , Austrosphaeromias , Neosphaeromias and Mackerrasomyia is the short female tergite 8, a feature also present in Palpomyiini + Stenoxenini . Further to this and discussed under character 134, there may be at least a short eversible sac present in Austrosphaeromias . In addition, the distinctive presence of setae restricted to the posteromedial margin of female sternite 8 is possibly a synapomorphy of Austrosphaeromias and some species of the distinta group of Palpomyia ( Spinelli et al. 2009b) . If the modifications to the male genitalia (character 115), tarsomeres 5 (character 101) and the claws (character 107) are deemed to be homoplastic, there is the possibility of the group being more closely related to the Palpomyiini + Stenoxenini or a group within Palpomyiini .

Mackerrasomyia is very similar to Neosphaeromias and the two genera may warrant synonymy ( Grogan & Wirth 1982). There are two synapomorphies combining the two: the presence of somewhat triangular coronal sutures in male Mackerrasomyia and both sexes of Neosphaeromias gibbus (but not in other species in this genus) (character 8 and somewhat homoplastic) and presence of a single spermatheca (character 146), unique within the lineage Heteromyiini + Hebetulini + Johannsenomyiini + Sphaeromiini + Palpomyiini + Stenoxenini other than Stenoxenini which also share this feature. Borkent (2017) keyed out Afrotropical males of these two genera based on relative length of the gonostylus but this feature does not distinguish all extraterritorial members of the genera (i.e. some Neosphaeromias have elongate gonostyli). Females were distinguished on the basis of the shape of internal teeth of the claws as well as the presence or absence of an external tooth on each claw. These features are known to be susceptible to homoplasy or loss respectively (see “Uncertain Character States” and character 107). A revision of all included species is needed.

It is important to recognize that the male and female of Dibezzia maybe not be associated correctly, as pointed out by Wirth & Ratanaworabhan (1981a) who noted differences in the relative length of the radial cells and distribution of palisade setae on the midtarsomere 1 (present in male, absent in female) and the hind leg (male with two rows on tarsomere 1 and a single row on tarsomere 2, female with 1 and a partial row only on tarsomere 1). However, differences in wing venation occurs in some other genera and reexamination of the female of Dibezzia debenhamae Wirth & Ratanaworabhan shows that in fact they actually have palisade setae on midtarsomere 1—therefore, the only difference between the sexes, regarding this feature is that male hind tarsomere 2 has the setae while these are absent in the female; this is likely a loss, considering that all other Sphaeromiini , Johannsenomyiini, Palpomyiini and Stenoxenini females have tarsomere 2 with palisade setae. Furthermore, it is notable that the only female Heteromyiini without palisade setae on hind tarsomere 2 are Heteromyia , Pellucidomyia and Tetrabezzia , also with females with extremely elongate hind legs. Another sexually dimorphic character in Dibezzia is that the males have eyes separated by 4 ommatidia while females have broadly abutting eyes although this sexual dimorphism is also present in a few other genera (character 2). Female Dibezzia have ventral katepisternal setae while these are missing in males (character 72); this feature is otherwise shared between the sexes in other genera with such setae.

The genus Lanatomyia , with three Australian species, is proposed here as the sister group of Johannsenomyia , with 27 species primarily in the Oriental, Australian and Afrotropical Regions and four in the New World. This relationship is based on three shared and unique synapomorphies: male hind tarsomere 5 with batonnets (character 102), male claws each with an external blunt tooth (character 103) and pupa with abdominal segments 3–6 with lateral setae L-2, L-3 and L-4 grouped on the anterior half of segment in Lanatomyia and further derived in Johannsenomyia with the setae L-2, L-3 and L-4 grouped on the anterior half of segment in a more or less transverse row (character 182; Borkent 2014, character 66). When Debenham (1974) proposed Lanatomyia , she indicated Lanatomyia differed from Johannsenomyia “in having the costa extending to the wing apex, only four pairs of batonnets which are restricted to the basal half of tarsomere 5, with the distal pair separated from the other three pairs, and the claws straight and stout with central swelling. The male parameres lack of lateral lobes characteristic of Johannsenomyia .” I consider the relative length of the costa to be a poor indicator of relationship (see in “Uncertain Character States”). The differences in the batonnets is very small. Many Johannsenomyia have 5 batonnets and in some, the distal pair is slightly separated from the more basal pairs, similar to the condition in Lanatomyia . Their claws are also similar (see below). The difference in the parameres is valid but it needs to be noted that intrageneric differences in paramere structure are present in many genera of Ceratopogonidae . Female Johannsenomyia have a small tooth between the two foreclaws. This feature is nearly unique in the Culicomorpha. Late in this study, I was able to study female Lanatomyia and they too have the feature, certainly a further synapomorphy uniting these two genera. The tiny tooth is difficult to discern in slide mounted specimens when the claws are in a lateral position. The feature is best seen in specimens in glycerine viewed dorsoventrally as well as some slide material with claws mounted in that position. The feature has not been previously reported in the literature. The following members of the Johannsenomyiini which were unavailable (or were in a poor orientation) were: Crispomyia , Groganhelea and Niphanohelea . This evidence indicates that future work should consider the possibility of combining the two genera, which are similar in other regards.

There are two character states which suggest that Crispomyia may be modified Nilobezzia . Crispomyia and some Nilobezzia have palpal segments 4+5 fused (character 44). Here it is presented as a synapomorphy of the two genera, with a reversal to the plesiomorphic condition in some Nilobezzia . Females with a medially divided abdominal tergite 9 is treated here as a synapomorphy of Niphanohelea + Nilobezzia + Crispomyia (character 140), with some reversals in Nilobezzia . There is a possibility of the Niphanohelea and Crispomyia are most closely related to only some Nilobezzia . The male genitalia of Nilobezzia is highly modified and it will be informative when the male of Crispomyia (and Niphanohelea ), presently unknown, is discovered.

As presented here, Palpomyiini is considered to be paraphyletic, with Phaenobezzia being the sister group of Stenoxenini . This is based on one synapomorphy: paratergite with setae (with independent losses in some Phaenobezzia and some Paryphoconus ) (character 66). I consider this to be weak evidence and therefore retain Palpomyiini until further study. The synapomorphy grouping Pachyhelea , Palpomyia , Bezzia and Phaenobezzia + ( Stenoxenus + Paryphoconus ) is the presence of dense setae on the scutum (character 60), a feature that occurs in some Johannsenomyiini as well and the plesiomorphic condition occurs in some Palpomyia , here treated as a reversal. Although Clastrieromyia , Pachyhelea and Phaenobezzia are each likely monophyletic, the whole tribe warrants further scrutiny, especially considering there is no synapomorphy indicating the monophyly of either Amerohelea or Palpomyia . The monophyly of Bezzia was indicated for a pupal synapomorphy by Borkent (2014: character 6) but this was missing in a couple of species of Bezzia and present in two species of Palpomyia . There is, therefore, the possibility that Bezzia or Palpomyia or both are paraphyletic in relation to any or all of the other genera of Palpomyiini .

Although Stenoxenus is clearly monophyletic, based on both male and female synapomorphies, the monophyly of Paryphoconus , its sister group, is less assured. The synapomorphies for Paryphoconus given here as autapomorphies and in the phylogenetic analysis (character 104, claw with strong inner tooth) are all of males. However, of the 41 species of Paryphoconus described, only seven are known as males and therefore there is uncertainty whether these synapomorphies apply to the entire genus. Spinelli & Wirth (1984: 883) discuss but do not describe a male which had the separate parameres of Paryphoconus but the characteristic four-segmented palpus of Stenoxenus and which they considered intermediate between Stenoxenus and Paryphoconus . I studied a second specimen (species?) with this same combination of features (also without an associated female). It had the anterior tentorial pit with a ventrally directed prong (an autapomorphy of Stenoxenus ), indicating that both fused and unfused parameres occur within Stenoxenus . This male is the one listed in Table 2 as Stenoxenus sp. At the present time, all described species of Paryphoconus are known only from the New World and nearly all only from the Neotropical Region. I have examined a single female from each of India (CAS; 24 km E. Trichur, Kerala) and Malaysia (CNCI; Pa Umor, Sarawak, Malaysia) which key to Paryphoconus other than them having a wing with two radial cells. Both specimens have the foretarsomere 5 markedly swollen, similar to those of Heteromyiini (see character 99) and represent a new genus of Stenoxenini .

In the section “Autapomorphies of Extant Genera”, the evidence for the monophyly of 67 genera is presented, generally as unique features in the family or even a broader group. In addition, the cladogram presents features which are autapomorphic for a given genus but these character states are all either reversals or homoplastic and these include 17 genera: Ankylohelea , Atyphohelea , Bothamia , Calcarhelea , Crispomyia , Homohelea , Mackerrasomyia , Metahelea , Neohelea , Notiohelea , Pachyhelea , Parabezzia , Physohelea , Probezzia , Pseudostilobezzia , Sphaeromias and African new genus A. Two genera have autapomorphies on the cladogram based solely on larval ( Sphaeromias ) or pupal features ( Parabezzia ); this is misleading in that the immatures of closely related genera, such Homohelea , Leehelea and Xenohelea in the case of Sphaeromias , and all other genera of Parabezziini in the case of Parabezzia , are unknown. Bezzia has a pupal autapomorphy ( Borkent 2014: character 6 but there were some exceptions and the feature is present in a few Palpomyia ). There are 24 genera which lack unique autapomorphies (some have homoplastic autapomorphies): Afrostilobezzia , Alloimyia, Amerohelea , Austrohelea , Brachypogon , Capehelea , Ceratohelea , Congohelea , Dibezzia , Downeshelea , Forcipomyia (but see autapomorphies for a possibility), Groganhelea , Indobezzia , Kolenohelea , Lanatomyia , Mallochohelea , Neurobezzia , Neurohelea , Palpomyia , Wannohelea , Washingtonhelea , Xenohelea , Yungahelea and Parabezziini new genus. For some of these, the genus is monotypic and therefore the question at this point is uninformative regarding their monophyly: Capehelea , Congohelea , Groganhelea , Neurohelea , Wannohelea , Washingtonhelea (a second undescribed species is known; CNCI) and Parabezziini new genus. However, when larger, more diverse genera lack evidence of monophyly, it brings into question whether they are paraphyletic and whether closely related genera are actually derived members of that larger genus. There are a number of areas of the phylogeny which in my opinion particularly require closer study, some of which are discussed elsewhere. The following discussion emphasizes some areas where evidence either indicates a related genus is paraphyletic in relation to another genus or, lacking evidence, is still a possibility.

In spite of the description of a potential autapomorphy for Forcipomyia , there remains a strong possibility that the genus is paraphyletic in relation to Atrichopogon . Clearly more research is needed of the many morphological differences, in every life stage, within Forcipomyia and Atrichopogon . Chinese taxonomists working on Forcipomyia continue to recognize the subgenus Forcipomyia (Lasiohelea) as a valid genus. There is some limited phenetic evidence indicating Atrichopogon and Forcipomyia (Lasiohelea) share some similarities, such as a more elongate costa (higher costal ratio), a fringe of simple setae on the posterior margin of the wing and unspecified antennal features ( Edwards 1922, Ingram & Macfie 1924). However, these features also occur in some other subgenera of Forcipomyia . Yu & Wirth (1997) provide a table comparing Forcipomyia , Atrichopogon and Lasiohelea (as a full genus) but clearly, the differences between the three are obscure. A cladistic analysis of the Forcipomyiini is needed to resolve their status.

Several features suggest that the very tiny adults of the monotypic Rhynchohelea , known only from a few females from Florida and California, USA and Costa Rica, may be merely specialized members of Brachypogon . The radial cells are all fused apically with the costa, similar to some Brachypogon (Brachypogon) . However, this feature occurs in some other Ceratopogonidae and is discussed in “Uncertain Character States” as “R 1 and R 3 joining costa as a strong, thickened fusion and with either no or one radial cell evident or with veins forming the second (or last) radial cell are swollen”. The reduced number of antennal flagellomeres of Rhynchohelea to 12 is also present in an undescribed species of B. ( Brachypogon ) from Thailand and Malaysia (pers. obs.); in addition, some African B. ( Brachypogon ) also have a reduced flagellum with 9–11 flagellomeres ( Grogan & de Meillon 1993). Finally, the female of the undescribed B. ( Brachypogon ) from Thailand and Malaysia and Rhynchohelea have an unusually high ratio of the fore- to hind tibia of more than 1.02, otherwise restricted to a group of Sphaeromiini (see character 85). This high ratio for Rhynchohelea and some Brachypogon is likely a synapomorphy, independently derived. In essence, the peculiarly developed mouthparts of female Rhynchohelea are their only distinguishing feature and as an autapomorphy could therefore have evolved within Brachypogon . The presently unknown male of Rhynchohelea will likely clarify its phylogenetic position. It is worth noting that the female of the undescribed species from Thailand and Malaysia has flagellomeres 2–4 tightly compressed and partially fused; the associated male has a much reduced flagellum, composed of the following flagellomeres: 1, 2–5 fused, 12, 13.

Afrostilobezzia is a genus with two West African species and known only as females. The only feature distinguishing them from Stilobezzia is that they have two instead of a single claw on each leg ( Szadziewski & Dominiak 2015). Clastrier (1988b) placed the only known species at that time within a framework of classification of Stilobezzia ( Clastrier 1976) . As discussed under “Uncertain Character States” as “Presence of one or two claws on each of female fore-, mid- and hind legs, each claw with or without basal or subbasal teeth” claw and tooth shape varies considerably and homologies are often uncertain. In this case, both equal and unequal claws are known within a few other genera: Serromyia , Alluaudomyia and Schizonyxhelea . It would not be unexpected to also occur in the diverse Stilobezzia (with 349 species worldwide). Further to this, the apparent two-clawed state of Afrostilobezzia may actually be merely a strongly developed inner tooth (shorter in Stilobezzia ); it is uncertain from the descriptions whether the “two claws” are fused basally or not ( Clastrier 1988b, Szadziewski & Dominiak 2015).

Austrohelea , without any autapomorphies, is likely composed of at least two unrelated lineages when undescribed species from Western Australia and New Zealand are considered. Further study is needed.

The monotypic Neohelea may be a derived Kolenohelea (see character 118).

Monohelea may be paraphyletic in relation to the monotypic Heterohelea . The male of Heterohelea has lost most of its wing pigmentation and has a reduced flagellum; a second, undescribed species has male wing pigmentation somewhat similar to that of some Allohelea (e.g. A. guineensis Clastrier & Delécolle ) (R. Szadziewski, pers. comm.).

Parabezzia needs revision, both in terms of undescribed species and their phylogenetic relationships. It may be paraphyletic in relation to Diaphanobezzia and, depending on the validity of the coxal spines (character 86, which varies within Parabezzia ), perhaps in relation to Spinellihelea and Yungahelea as well.

The relationships between the species of Probezzia , Mallochohelea + Neobezzia and Groganhelea may provide an improved understanding of the validity of these genera. Character 10 provides evidence that Neobezzia and Mallochohelea are sister groups but the feature was checked in very few species and the condition in Groganhelea is uncertain. As it is, they form part of a polytomy with another group of Johannsenomyiini .

As noted above in the discussion of Palpomyiini , the diverse genera in the tribe lack synapomorphies ( Fig. 9 ). Pachyhelea is a New World genus with two species. Grogan & Wirth (1980) discuss the similarities between this genus and the tibialis group of Palpomyia . It seems likely that when a cladistic study of Palpomyia is undertaken that Pachyhelea will be interpreted as an autapomorphic group of Palpomyia .

Conclusions and Suggestions for Future Research

One of most important goals of systematics is to provide phylogenetic interpretation which forms the basis not only of classification but as the predictive basis for interpretation of new taxa, newly discovered characters and newly discovered stages. Phylogeny also is the foundation for the logical interpretation of zoogeographic patterns, bionomic divergence and every other aspect of the organisms we study.

A clear conclusion from this paper is that the genera of Ceratopogonidae should be based on synapomorphies and phylogenetic interpretation (with broad outgroup comparisons). Previously, newly proposed genera were generally based on a mix of character states that were sometimes synapomorphic but were often homoplastic and/or not inclusive of all the species in the genus. Some genera have been based merely on a new combination of the features otherwise known in other genera. Peculiar and highly distinctive species need to be carefully scrutinized to determine whether they may be merely highly autapomorphic members of known genera, avoiding the temptation to provide a new name for an unusual species. Future descriptions of new genera will require the inclusion of many more adult and immature features than previously presented.

Because of the morphological divergence in the adult (both male and female), pupal and larval stages (and some eggs), the Ceratopogonidae are particularly amenable to searching for new synapomorphies in previously unknown stages, providing stronger and likely independent evidence from different semaphoronts (life stages) for phylogenetic relationships. When larval, pupal and adult synapomorphies are present on a particular node, it becomes increasingly certain that this reflects true historical genealogies.As such, the discovery of the immatures of genera where they are presently unknown ( Borkent 2014) will certainly be informative regarding the phylogenetic relationships of this family.

This study shows that some features previously utilized in the characterization of genera are markedly homoplastic or are puzzling in their character state distributions. Some of these may be more informative in the future when there is better understanding of these intragenerically. Regardless, it is important to point out that some of these features could be informative for identification and even if not present in all species in a given genus may yet be useful for the keying of genera, especially regionally. The goal of identification may not be the same phylogenetic resolution.

I take it as fundamental to our understanding of evolution that some characters are better indicators of genealogical relationships than others. Borkent (2018) indicated several means of judging which characters, a priori to their incorporation into a matrix, might be of greater weight than others. Part of that process of interpreting characters is the realization that some features are so homoplastic that they are almost certainly tied to particular and common environmental features. This is discussed more fully under “Phylogenetic Methods and Approach” above. As such, the study of more extensive outgroup comparisons will certainly provide critical information on the recognition of higher weight characters and a test of the phylogenetic relationships presented here. Further to this, some synapomorphies, treated as independent on a particular node may actually represent a single complex character (Borkent 2018). For example, my portrayal of three synapomorphies supporting the monophyly of the Ceratopogoninae other than the Culicoidini are all various modifications of the female mouthparts (characters 31, 35, 39) and perhaps represent a single and more complex evolutionary change. More synapomorphies that are clearly morphological unrelated on a single node provide the stronger evidence of monophyly.

Some instances of apparent homoplasy may be potentially teased apart by understanding their function. For example, the markedly elongate hind claws of the females of genera such as Heteromyia and Dibezzia , members of different tribes, may have fundamentally different roles in the behaviour of these taxa, supporting (depending on what is found) whether they are truly or only superficially homologous (see “Presence of one or two claws on each of female fore-, mid- and hind legs, each claw with or without basal or subbasal teeth” under “Uncertain Character States”). Resting females of Heteromyia and the closely related Tetrabezzia stand on their hind tibiae with their hind tarsi directly dorsally while those of Dibezzia stand with their hind tarsi on the substrate and the tip of the elongate hind claw touching the surface (based on photographs taken by Ian Jacobs on two separate occasions).

Study of the development of synapomorphies also may provide evidence of their particular weight, with complex development suggesting the likelihood of that feature being a strong indicator of relationship. It can also draw into question the validity of some evidence. Character 85, for example, reflects a unique leg feature of some pupae that was previously identified as a synapomorphy of Sphaeromias , Homohelea , Leehelea and Xenohelea . Some leg features of adults of these and one other genus, Indobezzia (unknown as a pupa) might have been scored as an independent synapomorphy if the connection had not been recognized that they were the same expression of a single feature in both pupa and adult. Similarly, the adults of Stilobezzia rabelloi and related species have an extended anterior protrusion of the scutum, a feature somewhat similar to a number of other Ceratopogonidae (see “Anteromedial tubercle on the scutum:” under Uncertain Character States). However, study of their pupae, which have highly modified pupal respiratory organs adapted to pierce the roots of aquatic plants, showed that the modification of the anterior of the adult thorax was an expression of the need in the pupa to orient and stabilize these unique pupal respiratory organs ( Borkent & Craig 2001). As such the modification evident in the adult is clearly independently derived from the scutal tubercles present in other Ceratopogonidae which lack the marked modification of the pupal respiratory organs. Further studies of the function of all the features of Ceratopogonidae could only lead to a better understanding of the relative strength of different synapomorphies.

In the quest to understand the evolution of our groups, discovering the function of some synapomorphies may provide evidence of their role in adaptive radiation and tying them to a particular and unusual behaviour. For example, the coarse serrate mandibles of female adults of the subfamily Ceratopogoninae other than Culicoidini (character 35) are associated in all known instances to their capture of small insects of similar size (almost always male chironomids), which they inject with a proteolytic enzyme and then ingest their liquified body contents. The coarse mandibular teeth certainly function to cut through the tough cuticle of their prey (compared to the finely toothed mandibles of vertebrate feeders). The feature may be further linked to the derived conditions of the labrum (character 31) and maxilla (character 39). These innovations of the mouthparts of Ceratopogoninae other than Culicoidini includes 94% of the generic diversity (101 genera) and 39% of species (2,455) of all known Ceratopogonidae ( Borkent & Dominiak 2020, Borkent et al. 2022). Similarly, the evolution of a particular larval form (character 187), a slender, in large measure naked (with fine, reduced setae), and flexible larva that moves snake-like through wet substrate, hunting in a very similar manner to snakes, probing here and there for prey, defines the subfamily Ceratopogoninae , including the above generic and species diversity of the family with the addition of Culicoides and Paradasyhelea and therefore the addition of 1,371 more species and including 61% of the species in the family.

The survey of adult characters here sometimes included only one or two species of a given genus (features only visible in glycerine specimens) and there is a strong need to expand comparisons within genera for these features. Furthermore, many genera are in need of revision and interpretation of their cladistic relationships. For example, for this study I set aside Macropeza blantoni Wirth & Ratanaworabhan females because of a number of differences distinguishing them from other members of the genus I examined (e.g. features of the wing and ommatidia). Future students of ceratopogonids undertaking generic revisions are strongly urged to include related genera to ensure that they are studying a monophyletic group (especially when related genera included few species).

This study includes only 16 larval synapomorphies, compared to 146 for adults and 40 for pupae. This reflects two issues. One is the paucity of comparative material. Of the larvae that have been preserved in museums, most are in poor to terrible condition, many compressed and many in an awkward orientation on slides. Old alcohol material is bleached and dehydrated, making it impossible to use these. The best work, by far, is that by Glukhova (1979) on the Russian fauna and most subsequent keys are based on her descriptions. Hribar & Mullen (1991) provided an excellent summary of previous work making comparisons between taxa, described the differences between the mouthparts of a variety of genera and gave detailed illustrations of each part. Second, it is striking that beyond the subfamily level, with the exception of the Forcipomyiinae , there are relatively few interpretable synapomorphies. The various stages of Forcipomyiini are remarkably rich morphologically and this has been utilized to recognize subgenera ( Saunders 1957); larvae can be identified to species in most cases (so far; the larvae of only 142 of the 1,146 species known have been described; Borkent 2014, Borkent & Dominiak 2020, Borkent et al. 2022). Although there has been no rigorous phylogenetic interpretation of the features of Forcipomyiini it is certainly a promising arena. However, in contrast to this, the larvae of the subfamily Ceratopogoninae , although moderately diverse morphologically, have few discrete features which can interpreted at the present time. Setae on the head vary substantially in length and position on the head capsule (but not many in relative position to one another) but I could see few interpretable features. The shape of the head capsule, shape and position of the developing eye, details of the pharyngeal complex, extent of the dorsoclypeal apotome anteriorly and other described features do not seem to be consistently different (or lack discrete character states). For example, the triangular or semispherical ventroposterior projection of the head capsule said to be characteristic of the Sphaeromiini (s. lat.) is not present in some species, or within some of the genera (e.g. Macropeza albitarsis Meigen ( Szadziewski & Dominiak 2007) , Probezzia seminigra (Panzer) ( Glukhova 1979)) and some are not notably different from some Palpomyia (e.g. P. tibialis ) (see character 191). Hribar & Mullen (1991) noted differences which seemed consistent between some groups. In addition, further outgroup comparisons are needed, particularly of the larvae. For example, the scoop-like mandibles of Leptoconopinae and Forcipomyiinae may be plesiomorphic based on these general forms of mandibles in Simuliidae and Thaumaleidae , but the mandible of Austroconops is slender, with adoral grooves ( Borkent & Craig 2004). The multidentate condition Hribar and Mullen (1991) noted as restricted to Forcipomyiinae is present in at least Leptoconops spinosifrons (Carter) ( Laurence & Mathias 1972) and likewise is present in both Simuliidae and Thaumaleidae . All the features noted by Hribar & Mullen (1991) need phylogenetic reappraisal, especially in the light of features of Austroconops (the generally conservative, earliest extant genus) and out group comparisons with other Culicomorpha. Hribar (1993) drew attention to the relationship between structure and function of the mouthparts of ceratopogonid larvae, which may, at least in part, correlate with phylogenetic diversification or perhaps explain some convergences. Further study is urgently needed, especially by comparing larvae in glycerine from a variety of angles and with SEM. The success of studies on the morphological differences of immatures of Simuliidae , Thaumaleidae, Culicoidea and Chironomidae (i.e. all other Culicomorpha) suggest that careful study of the larvae of Ceratopogoninae will be productive too.

As noted elsewhere immatures are known for only a limited number of genera, with pupae reported for 45 genera and larvae for 30 genera (out of 109 for the family). Although some of the genera without immatures may be because they are rare which is certainly true, for example, for some of the monotypic or otherwise small African and South American genera, there are some genera with numbers of species where the immature are unknown. Students of Ceratopogonidae need to keep in mind that these may be in unusual habitats not previously or poorly explored. For example, Echinohelea , a genus with 27 species, is known from one pupal exuviae reared from the margin of a pond in New York, USA. However, Grogan (1975) reported from a male reared from a pupa collected “from under the bark of a rotting tulip poplar ( Liriodendron tulipfer (sic))” in Maryland, indicating a fundamentally different habitat. When aware of the diversity of larval habitats present among species of Culicoides , known from, for example, fungi, rotting cacti, a wide array of aquatic habitats (trickles to lakes), phytotelmata, manure, rotting vegetation, red melting snow bearing algae, and more, it becomes increasingly possible that some of the genera with presently unknown immatures will also be found in what has been thought of as unlikely habitats.

So too, there is a great need for more material for many genera which are represented by so few adult specimens (sometimes 2–3) which are often in poor condition on microscope slide. This is particularly true for a number of African genera. With further material, better slide mounts could be prepared and adults could be studied in glycerine, allowing for the scoring of a number of character states which are presently unknown.

The Ceratopogonidae have a rich fossil record that has been very informative regarding the diversification of the family, representing one of the best fossil records in the Hexapoda. The fossil record reflects to a striking degree the cladistic relationships of the extant genera in the family. Many fossils, however, await description (e.g. at least 2,000 –3,000 specimens in Burmese amber housed in the Nanjing Institute of Geology and Palaeontology , Chinese Academy of Sciences, People’s Republic of China) and new deposits and further specimens from known deposits are being found nearly continuously. Some of these will provide further resolution regarding the generic phylogenetic relationships portrayed here as well as their temporal origins .

There has been little molecular evidence for phylogenetic relationships within Ceratopogonidae as discussed above under “Previous Phylogenetic Analyses”. Beckenbach & Borkent (2003) studied cox2 for 12 genera and showed some resolution for a few lineages. However, although initially meant to compare independent molecular and morphological results, some sequence analyses were tweaked in relation to the morphologically-based phylogeny to reach some congruence before publication.As suggested by Borkent (2018), this has happened for a number of other groups of Diptera . Strandberg & Johanson (2016) interpreted five genes for 32 genera. Their results produced some similarities to the phylogeny presented here but also with numerous differences as discussed above. Considering the continuing uncertainties regarding the accuracy of various molecular methods, it would be fascinating to generate a double blind study sequencing numbered but unnamed taxa and then compare this with morphologically-based analysis, which for the Ceratopogonidae is strongly predictive (especially the basal lineages).

It is notable that there are a significant number of genera of Ceratopogonidae with relatively few species, with 59 out of 109 genera (54%) known from five or fewer extant species (see Borkent & Dominiak 2020, Borkent et al. 2022). In the Holarctic Region, only Washingtonhelea (with a 2 nd undescribed species) and Atyphohelea are restricted to the Nearctic Region and two genera, Rhynchohelea and Pachyhelea , each with one species, occur in the Nearctic but with both also present further south in the Neotropics (both of these likely autapomorphic taxa of otherwise paraphyletic genera). Neurobezzia , with one species in the Nearctic is also known from a species in China and questionably, the Seychelles. In the Palaearctic Region, Neurohelea is the only certain endemic. There are a few small genera known from zoogeographically complex China with various definitions of the southern boundary of the Palaearctic Region (e.g. Holt et al. 2013, Xiaocheng et al. 2022); Wannohelea and Agilihelea are endemic in areas of China which could be considered Palaearctic. Similarly, Sinhalohelea and Pseudostilobezzia are also present in China and these genera also include species in the Oriental Region. The strong majority of small genera are from the Afrotropical, Neotropical and Oriental Regions. The sampling of the Afrotropical Region has been spotty. From 1956–1997 Clastrier published numerous species primarily from French speaking west Africa (e.g. Senegal, Guinea, Chad). In South Africa de Meillon published, as sole author in the early years of 1929– 1961 and from 1979–1991 with Wirth (see Borkent 2017), based on generally poorly preserved specimens and simplified illustrations. Excellent work completed by Spinelli and his associates in Argentina have opened up the diversity of that country and parts of Chile (including Patagonia) (see Spinelli et al. 2023). However, much of the Neotropical Region is unexplored. For example, the Andes of South America, one of most diverse areas on the planet for many other groups, are largely unsampled for Ceratopogonidae genera other than Culicoides . So too, the Australasian Region has a number of small genera, with the only broad taxonomic work largely done by Tokunaga and Debenham (see Debenham 1989). The Oriental Region also remains largely unexplored in spite of some fine work by Das Gupta and a more recent generation of researchers in India exploring their fauna. As such, it seems certain that numerous further species and likely further genera will be discovered in areas south of the Holarctic Region, especially when one considers the true relative size of countries and how little of these more southern areas, much larger than the Holarctic Region, have been collected.

I have not provided detailed zoogeographic analysis or discussion of bionomic divergence based on the phylogeny produced here. However, a few obvious zoogeographic patterns are worth noting here, as follows.

Genera in early lineages of Ceratopogonidae in Leptoconopinae and Forcipomyiinae are, or were, cosmopolitan in distribution. Austroconops is now restricted to southwestern Australia but fossils show that they and related genera (* Archiaustroconops , * Minyohelea , * Jordanoconops ) were broadly distributed in at least the Old World during the Cretaceous. Borkent & Craig (2004) suggested that the subsequent evolution of Culicoides (arising 99 mya) replaced Austroconops throughout most of its distribution.

The diverse genus Culicoides is nearly worldwide in distribution (not in New Zealand) and its sister group, Paradasyhelea is restricted to Gondwana (including New Zealand) and one species in the Pacific Northwest of the USA. The earliest fossil Culicoides ( Szadziewski et al. 2019a) are known from Burmese amber, about 99 mya, about the time that New Zealand (plus some associated islands such as New Caledonia) started to split from the rest of Gondwana ( Gibbs 2016), possibly explaining their present day distribution. Subsequent lineages are a mix of big genera which tend to be worldwide and smaller genera with more restricted distributions.Aside from Paradasyhelea (with the exception of one species in the Nearctic), two genera are restricted to Gondwana: Macrurohelea and Austrohelea (which may not be monophyletic).

Six genera forming a monophyletic group are restricted to the Afrotropical Region: Fanthamia , Afrohelea , Bothahelea , Ankylohelea , African new genus and Calcarhelea . The seven genera in Parabezziini are restricted to the New World, other than Parabezzia which is more broadly distributed (described species from the New World, Afrotropical Region, Sri Lanka; I have examined four undescribed species from Thailand (CSCA)).

There are further patterns of zoogeographic and bionomic interest but their interpretation awaits a future publication currently in preparation.

Figures

FIGURE 3. Thorax. A. Female adult Culicoides stigma, in anterior view (modified from Kaczorowska 2000, Fig. 27B). B. Female adult Sphaeromias pictus, in lateral view (modified from Kaczorowska 2000, Fig. 1, as S. candidatus (Loew)).

FIGURE 6. Phylogeny of the genera of Ceratopogonidae (second portion). Synapomorphy numbers not preceded by a letter are those of adults, “P- “indicates a pupal synapomorphy and “L- “a larval synapomorphy. Numbers which are underlined exhibit homoplasy and those in italics are reversals. Character state numbers coincide with the rectangle above them.

FIGURE 7. Phylogeny of the genera of Ceratopogonidae (third portion). Synapomorphy numbers not preceded by a letter are those of adults, “P- “indicates a pupal synapomorphy and “L- “a larval synapomorphy. Numbers which are underlined exhibit homoplasy and those in italics are reversals. Character state numbers coincide with the rectangle above them.

FIGURE 8. Phylogeny of the genera of Ceratopogonidae (fourth portion). Synapomorphy numbers not preceded by a letter are those of adults, “P- “indicates a pupal synapomorphy and “L- “a larval synapomorphy. Numbers which are underlined exhibit homoplasy and those in italics are reversals. Character state numbers coincide with the rectangle above them.

FIGURE 9. Phylogeny of the genera of Ceratopogonidae (final portion). Synapomorphy numbers not preceded by a letter are those of adults, “P- “indicates a pupal synapomorphy and “L- “a larval synapomorphy. Numbers which are underlined exhibit homoplasy and those in italics are reversals. Character state numbers coincide with the rectangle above them.

FIGURE 24. Frons and frontoclypeus of female Ceratopogonidae, in anterior view. A. Culicoides variipennis. B. Allohelea johannseni. C. Camptopterohelea n. sp. D. Cacaohelea glukhovae Borkent & Picado.

FIGURE 70. Genitalia of female Ceratopogonidae, in ventral view. A. Neobezzia costaricae. B. Mallochohelea smithi. C. Nilobezzia schwarzii. D. Pellucidomyia wirthi.

FIGURE 77. Stylized portrayal of female claw shape variation in Ceratopogonidae. Arrows show possible, logical directions of change, drawing into question homologous character states and making interpretation of synapomorphies difficult for most taxa.An exemplar taxon is noted for each form of claw shape and all extant Ceratopogonidae genera are listed in Table 5 under the corresponding figure letter. The variation is discussed under “Uncertain Character States” as “Presence of one or two claws on each of female fore-, mid- and hind legs, each claw with or without basal or subbasal teeth”.

FIGURE 78. A. Generalized wing of higher Ceratopogonidae based on Nilobezzia yasumatsui Wirth & Ratanaworabhan (from Wirth & Ratanaworabhan 1981b, Fig 2C). B. Theoretical evolutionary step showing distally displaced r-m. C. Theoretical evolutionary step showing r-m displaced further distally. D. Wing of Niphanohelea (from Grogan & Wirth 1981: Fig. 1A; their M1 and “adventitious vein” are interpreted here as r-m and M1, respectively).