The first fossil chylizine fly (Diptera: Psilidae) with a discussion on the ages of related fossil Acalyptratae

Abstract

The first fossil rust fly (Diptera: Psilidae) from the Insect Bed of the Bembridge Marls of the Isle of Wight is also the first known fossil record of the subfamily Chylizinae. It is late Eocene: Priabonian in age, 34.2 Ma. Chyliza colenutti Ross, Zhou, Hoffeins & Crighton, sp. nov. is compared with 30 extant species of Chyliza and differs from them in having a relatively long 2nd basal cell compared to the discal cell. Some other fossil Acalyptratae records are reviewed and their probable ages updated. The phylogenetic position of the Psilidae is currently uncertain so the family may have appeared anytime from 70–40 million years ago.

Muscomorpha, Schizophora, Diopsoidea, Chyliza, fossil, Bembridge Marls, Eocene, Priabonian

Introduction

Psilids (family Psilidae) are small-medium sized flies. Their larvae are phytophagous and feed on the stems and roots of various plants and a few species are agricultural pests, commonly known as rust flies (Evenhuis 2024). The family comprises about 340 described species world-wide and belongs to the superfamily Diopsoidea (Zhou and Yang 2022). 29 psilid species are found in Britain today (Hackston 2022). The family comprises three subfamilies: Belobackenbardiinae, Chylizinae and Psilinae, of which the first is regarded as the sister group to the rest of the family (Lonsdale 2020). The sister group relationship between the Chylizinae and Psilinae has been confirmed by molecular studies (Zhou and Yang 2022). The Chylizinae comprises only the one genus, Chyliza, the largest genus of the Psilidae, with 121 valid extant species (Systema Dipterorum, www.diptera.org).

Fossil psilids are poorly known with only three taxa previously recorded, however Gentilini et al. (2006) rejected Psilites bellaHeer, 1849 from the family and considered it may belong to the Tephritidae. Psila sp. identified by Schlöberlin (1888) from the Miocene of “Switzerland” requires confirmation (Evenhuis 1994). The only hitherto reliable fossil record of this family is Electrochyliza succiniHennig, 1965 from Baltic amber, however at least one undescribed species in Baltic amber is also known according to Tschirnhaus and Hoffeins (2009). Electrochyliza is unplaced to subfamily and is phylogenetically placed as the sister group to Chylizinae + Psilinae (Lonsdale 2020) (Fig. 1). The new fossil, from the Isle of Wight, UK, is the second reliable fossil psilid species to be described and provides an accurate calibration point for phylogenetic studies.

Fig. 1.

Phylogeny of the subfamilies of Psilidae following Lonsdale (2020), plotted against part of the geologic time scale, with fossil records. O = Öhningen, BM = Bembridge Marls, BA = Baltic amber, 1 = Electrochyliza succini, 2 = Psila? sp., 3 = Chyliza colenutti.

The new fossil psilid is preserved in Insect Limestone, a fine-grained micrite found in the Insect Bed of the Bembridge Marls Member, Bouldnor Formation of the Isle of Wight, southern England. The limestone has exquisitely preserved the insects, often in 3D, and has been referred to as “opaque amber.” It is of late Eocene: late Priabonian age and is 34.2 million years old (Ross and Self 2014). Fossil Diptera from the Bembridge Marls were studied for the INTAS funded project The terrestrial fauna and flora of the Insect Bed, Isle of Wight: interpreting the climate near the Eocene/Oligocene boundary (Ross 2014). The Acalyptratae were studied by Zlobin (2007), the Nematocera were studied by Krzemiński et al. (2019), except for Sciaridae which were studied independently by Greenwalt et al. (2019), and the Stratiomyidae were studied by Nicholson in Krzemiński et al. (2019), though other families of Brachycera were not studied.

Only six species of Acalyptratae have been previously named from the Bembridge Marls, described by Cockerell (1915, 1917), and their family placements have changed with time. Evenhuis (1994) listed the species under the families Chloropidae, Ephydridae, Richardiidae, and Sphaeroceridae. Zlobin (2007) reviewed these records and studied additional specimens at the Natural History Museum, London. He moved Protoscinus perparvusCockerell, 1917 from the Chloropidae to the Ephydridae; Ephydra oligocenaCockerell, 1915 and E. sepultaCockerell, 1915 from the Ephydridae to the Anthomyzidae; Stenomyites fuscipennisCockerell, 1915 from the Richardiidae to the Otitidae, and Sphaerocera sepultaCockerell, 1915 from the Sphaeroceridae to the Ephydridae. Additionally, Zlobin recognized the families Agromyzidae, Heleomyzidae, and Lauxaniidae but did not describe any new species. Roháček (2012) subsequently rejected “Ephydra” oligocena and “E”. sepulta from the Anthomyzidae and Nicholson et al. (2015) incorrectly listed "Sphaerocera"sepulta under the Sphaeroceridae. Krzemiński et al. (2019, table 3) listed all the taxa mentioned in Zlobin (2007) but missed Roháček (2012)’s paper. The Otitidae are now included in the family Ulidiidae (see Han and Ro 2016), so Stenomyites fuscipennis and additional undetermined specimens identified as belonging to Otitidae by Zlobin (2007) are hereby moved to the Ulidiidae. The family Psilidae has not been previously recorded from this fauna, thus this fossil constitutes the first record of this family from the Bembridge Marls. See Table 1 for a list of the known Acalyptratae from the Bembridge Marls.

Table 1.

List of Acalyptratae recorded from the Insect Bed of the Bembridge Marls, Isle of Wight

Family and subfamily	Genus and species
Agromyzidae: Agromyzinae	Undetermined
Agromyzidae: Phytomyzinae	Undetermined
Anthomyzidae	Undetermined
Chloropidae	Hippelates brodieiCockerell, 1915
Chloropidae	Undetermined
Ephydridae: Ephydrinae	Protoscinus perparvusCockerell, 1917
Ephydridae	“Sphaerocera” sepultaCockerell, 1915
Heleomyzidae	Undetermined
Lauxaniidae	Undetermined
Psilidae: Chylizinae	Chyliza colenuttisp. n.
Ulidiidae	Stenomyites fuscipennisCockerell, 1915
Ulidiidae	Undetermined
Incertae sedis	“Ephydra” oligocenaCockerell, 1915
Incertae sedis	“Ephydra” sepultaCockerell, 1915

The Psilidae have a characteristic wing venation, which, following the terminology used in Lonsdale (2020, following Cumming and Wood 2017), comprises a break in the Costa (C) situated from about 1/4 to over 1/3 the length of the wing, called the subcostal break. The break allows the wing to bend down towards the abdomen when resting. The weak Subcosta (Sc) runs very close to the Radius (R) then connects to C immediately before the costal break, delimiting the subcostal cell (though the connection is not always visible), and the humeral cross-vein connects C and Sc near the base of the wing. R bifurcates into R1 and the Radial sector (Rs) near the base of the wing. R1 curves upwards to merge with the Costa (C) before or at the middle of the wing. Rs bifurcates into R2 + 3 and R4 + 5 from nearly 1/4 to 1/3 the length of the wing. The Median (M) bifurcates into M1 and M4 near the base of the wing and there is a short cross-vein (r-m) between R4 + 5 and M1 situated from the basal 1/3 to the middle of the wing. M1 and M4 are connected by two cross-veins (bm-m and dm-m) which delimit the 2nd basal cell and large discal cell respectively. The bm-m cross-vein and the bifurcation of Rs are equidistant from the base of the wing. M4 and CuA are fused at the base and CuA forms a cross-vein between M4 and CuP, which delimits the anal cell. CuA + CuP continues posteriorly but terminates before reaching the hind-margin.

A useful wing character, which can be used to separate Chyliza from other psilid genera, is that Chyliza has a distinctively shorter anal cell compared to the 2nd basal cell (Hackston 2022; though it can be slightly shorter in other taxa). This can be clearly seen in the fossil specimen thus it can be confidently identified as belonging to Chyliza and constitutes the first fossil record of this genus and the subfamily Chylizinae. By contrast Electrochyliza has the two cells of equal length (see Hennig 1965, fig. 67 and Fig. 2), more typical of members of the Psilinae.

Fig. 2.

Electrochyliza succini in Baltic amber. Hoffeins Collection CCHH 1829-10. Scale bar 1 mm.

Given that the Psilidae has a very poor fossil record and a large number of extant species, this implies that this family has diversified rapidly relatively recently at the species level. It could be argued that because these are small flies they are hard to recognize as fossils, however this family is notably absent from ambers that are younger than Baltic amber (e.g., Dominican and Mexican amber) and would be expected to be present if the family had diversified earlier. Although the fossil can only be described from wing characters, the chance of this 34.2 million-year-old fossil belonging to an extant species is extremely remote, particularly as no adults of any extant Diptera species have been confirmed as present in the pre-Quaternary fossil record (before 2.58 Ma) (see Catalog of the fossil flies of the world, http:/hbs.bishopmuseum.org/fossilcat). Older galls have been attributed to extant species of cecidomyiid (see Skuhravá and Skuhravý 2010), but it is possible they were made by extinct species. The previous oldest record of an extant species, Fannia scalaris (Fabricius, 1794), recorded by Hennig (1966) from Baltic amber, was found to be a fake (Grimaldi et al. 1995, Ross 2010).

The Chyliza fossil has been compared to the wings of 30 extant species (see below) and there are notable differences, thus this fossil can be described as a new species, Chyliza colenuttisp. nov.

Materials and Methods

The specimen, in two parts, was found in a small collection of fossil flies and other insects from the Bembridge Marls collected by George William Colenutt and deposited at the Oxford University Museum of Natural History (OUMNH) in 1931 (Poulton in Marshall and Staley 1931; Oxford University Museum Annual Report for 1931, Eliza Howlett pers. comm. 2023). Colenutt probably collected it in 1929 at Gurnard Point (=Gurnard Ledge), north-west Isle of Wight, where he is known to have found fossil insects (Davis 1945). In 2022 the opportunity to study the OUMNH Colenutt Collection became available, thanks to funding from the National Science Centre, Poland.

The fossil was incorrectly identified as belonging to the Stratiomyidae and an initial comparison of the fossil with published papers on fossil and recent Diptera revealed that it was similar to Chyliza notata as figured in Lonsdale (2020, fig. 396). Extant species of Chyliza are distinguished based on coloration of the head and thorax, and features of the genitalia, which unfortunately are not visible in the fossil. The fossil has well-preserved wings but unfortunately in the primary descriptions of extant species the wings have been largely ignored, with only brief inconsistent descriptions which are not useful for detailed comparison. The fossil was visually compared to extant Chyliza specimens at National Museums Scotland, Edinburgh (NMS), The Natural History Museum, London (NHMUK), China Agricultural University, Beijing (CAU), and published photos and line drawings of other species. For some specimens, not all wings and not all characters were visible if a wing was missing, broken, bent, crumpled, or obscured; and for some species at the NHMUK represented by many specimens, not all of the specimens were studied. In total the fossil was compared to 177 wings of 30 extant Chyliza species. During this study it was realized that the “type” of Chyliza cylindrica (Walker, 1853a) had been misidentified and the correct name for this species should be C. pallidipesLamb, 1917 (see Ross et al. 2024).

This paper and the nomenclatural act it contains have been registered in Zoobank (www.zoobank.org), the official register of the International Commission on Zoological Nomenclature. The LSID (Life Science Identifier) number of the publication is: urn:lsid:zoobank.org:pub:DC15F584-33C9-40FE-B7CA-5FC5C206A612.

Results

To compare the fossil with the wings of all known 121 extant species of Chyliza is beyond the scope of this study and would be extremely challenging given that for most of the species wing images have not been published and detailed wing descriptions are lacking. However, it has been possible to compare it with specimens, photos and line drawings of 30 extant species (Fig. 3). Initially it was considered that there were several characters that could be used to distinguish the fossil from extant species, but from close examination of the NMS and NHMUK collections it was revealed that some characters were variable, or it was not possible to take some accurate measurements from photos as there was uncertainty as to whether the wing was completely flat. Ideally landmark morphometrics is required for the comparison of detached flattened wings, but that is beyond the scope of this study. Although the wing length of type specimens is often provided in primary descriptions, the size of the wing varies in a species. For instance, in only four specimens of C. vittataMeigen, 1826 in the NMS collection the wing length was seen to range from 4.2 to 5.8 mm and in eight specimens of C. munda (Walker, 1860) in the NHMUK collection the wing length ranged from 3.3 to 5.0 mm. The length/width ratio was examined, however if the wing is not flat then it can appear narrower. It was also considered that there may be differences in the distance from R1 to the costal margin in the subcostal cell compared to the distance from R1 to R2 + 3, however if the wing is not flat then the latter can appear narrower. Similarly, comparing the lengths of the bm-m and CuA cross-veins can also be unreliable if the wing is not flat, though in examined specimens they were seen to be of similar length, as also seen in the fossil. The fossil has an evenly curved hind-margin which is similar to some of the species; others have an emarginated hind-margin where M4 and CuA + CuP terminate, though this is accentuated if the wing is not flat. Shatalkin (1998, 2014) provided relative lengths of three sections of M1 for 28 species however these lengths are variable with differences seen on the left and right wings of individuals. E.g. for C. munda it was stated “Section of M_1 + 2 between r-m and dm-cu about 1.8 times more than previous one and about 1.4 times less than ultimate one” (Shatalkin 1998, p. 105). However, in the left wing of the Lectotype at the NHMUK, the section between r-m and dm-m is 1.5x more than the previous section (to bm-m) and the ultimate section (to the wing margin) is 1.6x more than the r-m to dm-m section (the right wing is missing). In the Paralectotype left wing, the r-m to dm-m section is 2.2x more than the previous section and the ultimate section is 1.3x more than the r-m to dm-m section. In the right wing it is 2.0x and 1.2x more respectively.

Fig. 3.

Example of an extant species of Chyliza for comparison. Chyliza callosa NHMUK Paratype. Its wing differs from C. colenutti in having a longer R1 compared to the distance of r-m from the base, a kink in M1 before the bm-m cross-vein, the 2nd basal cell being half the length of the discal cell, and an emarginated hind-margin. The bm-m cross-vein appearing shorter than the CuA cross-vein is because the wing isn’t flat.

Three characters were found for which the fossil differed from most or all examined extant species (Table 2). The first is that the distance from the base of the wing to where R1 terminates is the same as the distance to the r-m cross-vein. For most extant species R1 is longer, though is shorter in Chyliza leguminicolaMelander, 1920. The distance is the same for four extant species, however for five other species the distance was either the same or R1 was longer. The second character is that there is often a kink in M4 before the bm-m cross-vein in extant species but is straight in the fossil. It was also seen to be straight in three extant species, however in three other species it was seen to vary from straight to kinked. The third character distinguishes C. colenutti from all the examined extant species and that is the relative length of the 2nd basal cell compared to the discal cell (measured along M4), which is significantly longer in the fossil than in the extant species, being 0.7x the length of the discal cell. All the examined extant species were in the range 0.4–0.6x, with most 0.5x.

Table 2.

The main differences between Chyliza colenuttisp. n. and 30 extant species based on the comparison with specimens, photos, and line drawings. (H) = Holotype, (L) = Lectotype, (Pl) = Paralectotype, (N) = Non-type, (ld) = line-drawing, (P) = Paratype, (Ps) = Paratypes, (Ss) = Syntypes; “R1 v. r-m” compares the distance of R1 and the r-m cross-vein from the base of the wing; “M1” is whether M1 is straight or kinked just before the bm-m cross-vein, and “2nd bc v. dc” is the relative length of the 2nd basal cell compared to the length of the discal cell

Species	Type?	Source, number of wings examined	R1 v. r-m	M1	2nd bc v. dc
†C. colenuttisp. nov.	H	OUMNH, 2 wings	equal	straight	0.7x
C. acromelaena	N	NHMUK, 8 wings	equal	slightly kinked	0.5x
C. annulipes	N	NHMUK + Lyneborg (1964), 24 wings	longer	straight to kinked	0.5–0.6x
C. apicalis	N ld	Steyskal (1987), 1 wing	equal	kinked	0.5x
C. bambusae	P + N	CAU, 2 wings	longer	kinked	0.5x
C. callosa	H + P	NHMUK, 4 wings	equal or longer	kinked	0.5x
C. caudata	Ps	NHMUK, 4 wings	longer	kinked	0.5–0.6x
C. chikuni	N	CAU, 1 wing	longer	slightly kinked	0.4x
C. coxachaeta	P	NHMUK, 2 wings	longer	kinked	0.5x
C. extenuata	N	NMS, NHMUK, 19 wings	equal or longer	kinked	0.5–0.6x
C. flavicrura	H	CAU, 1 wing	longer	kinked	0.5x
C. fumipennis	H ld	Morge (1980), 1 wing	longer	kinked	0.5x
C. huiana huiana	H	CAU, 1 wing	equal	kinked	0.5x
C. ingetiseta	H + N	CAU, 2 wings	longer	kinked	0.5x
C. leguminicola	N	Bygebjerg et al. (2011) + Lonsdale (2020), 2 wings	shorter	slightly kinked	0.5x
C. leptogaster	N	NMS, NHMUK 16 wings	longer	straight to kinked	0.4–0.5x
C. maculasterna	P + N	CAU, 2 wings	longer	kinked	0.4–0.5x
C. maculifrons	H	NHMUK, 2 wings	longer	slightly kinked	0.5x
C. munda	L, Pl + N	NHMUK, 13 wings	equal or longer	kinked	0.5–0.6x
C. notata	N	Lonsdale (2020), 1 wing	longer	straight	0.6x
C. nova	N	NHMUK, 16 wings	longer	straight	0.5x
C. pallidipes	H + N	NHMUK, 11 wings	equal or longer	kinked	0.4–0.6x
C. pictipennis	H + P	NHMUK, 4 wings	longer	kinked	0.5–0.6x
C. prominens	H	NHMUK, 2 wings	longer	slightly kinked	0.5–0.6x
C. qaradaghi	H	Khaghaninia & Gharajedaghi (2014), 1 wing	equal	straight	0.5x
C. rwindiensis	N	NHMUK, 2 wings	longer	kinked	0.5x
C. selectoides	N	NHMUK, 2 wings	longer	slightly kinked	0.5x
C. sinensis	P	CAU, 1 wing	longer	kinked	0.5x
C. varipes	Ss	NHMUK, 9 wings	longer	kinked	0.5–0.6x
C. vittata	N	NMS, NHMUK + Sueyoshi (2013), 20 wings	equal or longer	straight to kinked	0.4–0.5x
C. zangana	H	CAU, 1 wing	longer	kinked	0.4x

Systematics and Taxonomy

Order Diptera Linnaeus, 1758

Suborder Brachycera Zetterstedt, 1842

Infraorder Muscomorpha Crampton, 1944

Section Schizophora Becher, 1882

Subsection Acalyptratae Macquart, 1854

Superfamily Diopsoidea Billberg, 1820

Family Psilidae Walker, 1853b

Subfamily Chylizinae Rondani, 1856

Genus ChylizaFallén, 1820

Chyliza colenutti sp. nov.

Figs 4–6

Fig. 4.

Chyliza colenuttisp. nov., Holotype; A) part, OUMNH M.177; B) counterpart, OUMNH M.175; Insect Limestone, Bembridge Marls, north-west Isle of Wight, UK. Colenutt Collection. Scale bar 5 mm.

Fig. 5.

Chyliza colenutti, Holotype, close-up of wings; A) right wing of part; B) left wing, as preserved on the right side of the counterpart, with negative relief; scale bar 1 mm.

Fig. 6.

Chyliza colenutti, Holotype, line-drawing showing wing venation. C = Costa, Sc = Subcosta, R = Radius, M = Median, Cu = Cubitus.

urn:lsid:zoobank.org:act:775D51C7-8A6F-437C-BF37-EA1EF3EAF947

Diagnosis

Consistent with Chyliza in having a distinctively shorter anal cell compared to the 2nd basal cell in the wing. The new species has the combination of R1 terminating on the costal margin at the same distance along the wing as the r-m cross-vein, M1 straight before the bm-m cross-vein, and a relatively long 2nd basal cell (0.7x the length of the discal cell).

Description

Preserved body length 3.9 mm, thorax width 1.1 mm, wing length 3.6 mm, wing width 1.4 mm, abdomen width 1.0 mm. Wing length/width ratio 2.6. Trichia clearly visible along the anterior margin of the wing. The wings are pigmented in the anterior distal region of the wing, and the hind-margin is evenly curved (not emarginate where M4 and CuA + CuP terminate). The distances between R1 and the costal margin in the subcostal cell and R1 to R2 + 3 are similar. Cross-veins bm-m and CuA are of similar length.

Holotype

OUMNH M.177 (part) and M.175 (counterpart), sex unknown, Colenutt Collection.

Etymology

Named after George William Colenutt who collected the specimen.

Locality and Horizon

Gurnard Ledge, north-west Isle of Wight, UK [50.7498°N, -1.3453°W]; Insect Limestone, Bembridge Marls Member, Bouldnor Formation; late Eocene: Priabonian.

Discussion

The occurrence of Chyliza colenuttisp. nov. is significant because it provides a useful calibration point within the phylogeny of the family, demonstrating that Chylizinae and Psilinae diverged before 34.2 Ma. The family Psilidae must have appeared before then, but when?

According to the phylogeny in Lonsdale (2020), the Psilidae comprise the sister group to the rest of the superfamily Diopsoidea. The remaining families include Nothybiidae, Somatiidae, Gobryidae, Diopsidae, Megamerinidae, and Syringogastridae, of which the last three families were well-supported as sister-families and are the only families confidently placed in the superfamily. However, the monophyly of the Diopsoidea was considered weak by Lonsdale (2020) and has been disputed in molecular studies. Bayless et al. (2021) recovered Psilidae as the sister group to the Opomyzidae, whereas Zhou and Yang (2022) suggested a relationship of Psilidae with either Agromyzidae (Bayesian inference), or Diopsidae + Nothybidae (maximum likelihood).

Having accurate ages of first fossil records is important for their use as calibration points for phylogenetic studies to estimate the time of origin of taxonomic groups. The probable age of Chyliza colenutti is quite precise, however the interpreted ages of many other fossil insects are often imprecise, particularly those from non-marine deposits that are hard to correlate. Interpretations of sedimentary deposits change as further geological research, dating techniques, and correlation with other deposits refines their age. Simultaneously, the dating of international chronostratigraphic boundaries is also being refined (Gradstein et al. 2020). It is also worth bearing in mind that old localities may now be in a different country due to boundaries moving. For instance, the classic Miocene locality of “Oeningen in Switzerland”, where Psila? sp. came from, today is Öhningen in southwestern Germany (Salvador et al. 2022). Fikáček and Schmied (2013) referred to the fossil insects from this locality as coming from the Upper Freshwater Molasse, in the European Land Mammal Mega Zones MN6-7, which correlate to the Middle Miocene: late Langhian to Serravallian stages, about 15–12 Ma (Hilgen et al. 2009), so from taking a mid.-point the fossil record of Psila? sp. is about 13.5 million years old.

Many first fossil records of extant insect families are found in amber, e.g., Electrochyliza succini and undescribed psilid species from Baltic amber constitute the oldest records of Psilidae, but the dating of amber is problematic in that it can only be done by dating the bed that the amber is found in. The age of Baltic amber is notoriously imprecise as it comes from different horizons (Standke 2008), and most specimens come into scientist’s hands via the amber trade, after precise information on their locality and horizon are lost (though for pieces washed up on beaches that information is already lost). It ranges from the middle Eocene to middle Oligocene (48 to 27 Ma), however most mined amber came from the Upper Blue Earth Member of the Prussian Formation and is the most likely source of the amber washed up on Baltic beaches. This member has often been referred to as middle Eocene: Lutetian in age based on the radioisotopic dating of glauconite from the sediments (Ritzkowski 1997), however this technique is not reliable as if the sample is not pure then contaminants can provide an incorrect older date (Clauer et al. 2005, Wu et al. 2023). Biostratigraphical correlation places the Upper Blue Earth within the late Eocene: Priabonian stage (Kasiński et al. 2020, see Grimaldi and Ross 2017, appendix 1 for a review of the ages of Baltic and other ambers). The Priabonian ranges from 37.71 to 33.9 Ma according to the latest International Chronostratigraphic Chart, v2023/09 (https://stratigraphy.org/chart). The identification of microfossils placed the Upper Blue Earth within dinoflagellate cyst zone D12 (Kasiński et al. 2020), which ranges from 36.5 to 33.5 Ma (Spiejer et al. 2020), so a mid-point of 35 Ma can be used as an approximate age for Baltic amber. There are 37 families of Acalyptratae recorded from Baltic amber (Michelsen 2009, Tschirnhaus and Hoffeins 2009, Nicholson et al. 2015, Barták 2019, Lonsdale 2020, Roháček and Hoffeins 2021, Roháček et al. 2023). This high diversity indicates there was an “explosive” radiation of the Acalyptratae in the Eocene (Roháček et al. 2023).

Regarding the other families of Diopsoidea, sensuLonsdale (2020), Nothybiidae, Somatiidae, and Gobryidae do not have a fossil record. The oldest fossil Diopsidae are three described and one undescribed species from Baltic amber, of which one is also known from Rovno amber of the same age (Priabonian) (Perkovsky et al. 2015, Feijen and Feijen 2023), whereas Prosphyracephala rubiensisLewis, 1971 from the Medicine Lodge Formation of the Ruby River Basin, Montana, USA is younger, of Early Oligocene: Rupelian age (33.9–27.8 Ma), about 31 Ma (Stigall et al. 2017). The Megamerinidae is represented by one described species and one undescribed species from Baltic amber (Tschirnhaus and Hoffeins 2009), and the Syringogastridae by two species from Dominican amber of Early Miocene: Burdigalian age (20.44–15.98 Ma), about 18 Ma (Marshall et al. 2009). In Lonsdale’s (2020) phylogeny, the Diopsidae are the most derived family of the Diopsoidea, so given that the oldest fossil diopsids are Priabonian in age then the other diopsoid families probably appeared before then. So, if Psilidae is the sister group to the rest of the Diopsoidea then this family probably appeared much earlier. Conversely, Bayless et al. (2021) recovered Psilidae as the sister group to the Opomyzidae. Only two fossil opomyzids are known, which came from the Rott Formation, Rott, Germany (Stadtz 1940), in mammal zone MP 30, near the end of the Chattian Stage (Late Oligocene), 23.1 Ma (Kotov and Wappler 2015, Spiejer et al. 2020). They are younger than Chyliza colenutti and Electrochyliza, which still indicate that the Psilidae appeared before the Priabonian but may not have appeared long before. Using Bayesian inference Zhou and Yang (2022) suggested the Psilidae are closely related to the Agromyzidae. The Agromyzidae has a good fossil record with many Cenozoic records, based on adults and trace fossils (Winkler et al. 2010). Although this family has been recorded from the Bembridge Marls, it is not known from Baltic amber (previous records have been moved to other families). Three species based on body fossils have been described from the Florissant Formation of Florissant, Colorado (Melander 1949, Meyer 2003) of similar age to the Bembridge Marls. A radiometric date of 34.07 Ma was taken from near the top of the Florissant Formation and the formation correlates to the lower part of the C13r magnetostrat chron (Evanoff et al. 2001, Prothero and Sanchez 2004). The base of this chron is dated as 35.1 Ma (Spiejer et al. 2020), so the Florissant Formation ranges from 35.1 to 34.0, with a mid-point of about 34.5 Ma.

The oldest undoubted fossil acalyptrate fly belongs to the Lauxaniidae from French Oise amber of early Eocene: Ypresian age, about 53 Ma (Roháček and Hoffeins 2023). However, the oldest evidence of Acalyptratae are fossil leaf-mines attributed to the Agromyzidae, named Phytomyzites biliapchaensisWinkler et al. 2010 from Mexican Hat, Montana, USA, of early Palaeocene: early Danian age, about 65 Ma (Donovan et al. 2014), only one million years younger than the end-Cretaceous extinction. Their age fits well with the phylogeny of Junqueira et al. (2016) who placed Liriomyza (Agromyzidae) as originating near the base of the Schizophora tree, which could imply that recovery after the extinction provided the opportunity for the Schizophora to diversify. However, Junqueira et al. (2016) did not include psilids in their study and incorrectly referred to the Agromyzidae as occurring in Baltic amber. Conversely, the molecular studies of Bayless et al. (2021) and Zhou and Yang (2022), which place the Agromyzidae well within the Schizophora tree, may imply that the Schizophora radiation started within the Late Cretaceous.

Given that the phylogenetic position of the Psilidae is currently uncertain (either sister group to the rest of the Diopsoidea, or to the Opomyzidae or even Agromyzidae), then the family could have appeared anytime from about 70 to 40 million years ago. Further study is required integrating molecular, morphological, and fossil data.

Conclusions

The discovery and study of a fossil fly from the Bembridge Marls of the Isle of Wight in the collections of the OUMNH has led to the description of Chyliza colenuttisp. nov. It can be placed in the extant genus Chyliza with confidence and can be demonstrated to be different from 30 extant species of Chyliza. The new species constitutes the first record of the family Psilidae from the Bembridge Marls and the first fossil record of the subfamily Chylizinae. It forms a useful calibration point for phylogenetic studies, demonstrating that the Chylizinae, and Psilinae diverged before 34.2 million years ago and that the family Psilidae appeared sometime before then.

A review of some related fossil Acalyptratae records has provided more accurate ages which could be useful for calibrating further phylogenetic studies.

Version of Record, first published online November 13, 2024 with fixed content and layout in compliance with Art. 8.1.3.2 ICZN.

Disclaimer: Nagoya Protocol: No specimens were included in this project that required permits of any kind.

Specimen Collection Statement

The authors attest that all legal and regulatory requirements, including export and import collection permits, have been followed for the collection of specimens from source populations at any international, national, regional, or other geographic level for all relevant field specimens collected as part of this study.

Acknowledgements

The authors thank Owen Lonsdale and an anonymous reviewer for constructive comments and suggestions. AJR thanks Eliza Howlett (OUMNH) for making Colenutt’s collection available, providing information on when the collection was acquired and for lending the specimen; to Milo Phillips and Erica McAlister for access to the extant psilid collections at NMS and NHMUK respectively, to Vladimir Blagoderov (NMS) for supplying literature, and to the National Science Centre, Poland for funding this research, project No. UMO-2020/37/B/NZ8/03042.

Author contributions

Andrew Ross (Conceptualization [lead], Investigation [equal], Methodology [lead], Resources [equal], Writing—original draft [lead], Writing—review & editing [lead]), Jiale Zhou (Conceptualization [supporting], Investigation [equal], Methodology [supporting], Resources [equal], Writing—original draft [supporting], Writing—review & editing [supporting]), Christel Hoffeins (Conceptualization [supporting], Investigation [supporting], Methodology [supporting], Resources [equal], Writing—original draft [supporting], Writing—review & editing [supporting]), and Bill Crighton (Resources [equal], Writing—review & editing [supporting])

Funding

This study was supported by National Science Centre, Poland, Grant No. UMO-2020/37/B/NZ8/03042.