Morsagrion Zessin, 2011: 65

Archibald, S. Bruce, Ware, Jessica L., Rasmussen, Jan A., Sylvestersen, René L., Olsen, Kent & Simonsen, Thomas J., 2023, The damselfly genus Furagrion Petrulevičius et al. (Odonata, Zygoptera) from the early Eocene Fur Formation of Denmark and the dysagrionoid grade, Zootaxa 5278 (2), pp. 289-317 : 292-302

Morsagrion Zessin, 2011: 65 .

Type species: Morsagrion ansorgei Zessin, 2011 , by original designation. Syn. nov.

Figs. 1–4 , 6–8 , 1124.

Included species. Furagrion jutlandica ( Henriksen, 1922) , Furagrion ansorgei ( Zessin, 2011) comb. nov.

Range and age. Earliest Ypresian Fur Formation, Jutland, Denmark. Specimens of this genus are known to range from just below ash layer -13 within the Knudeklint Member to the lower part of the Silstrup Member (the ash layer +25 – +30 interval) of the Fur Formation, which has an absolute age of approximately 55.5 Ma ( Storey 2007, Stokke et al. 2020). Six of the sixteen examined fossils, however, are from unknown levels within the Fur Formation.

Emended diagnosis. The wings of Furagrion differ from those of similar extant and extinct Zygoptera by possession of all nine wing character states of the Dysagrionidae Cockrerell (Cephalozygoptera) diagnosis listed below, but the genus is excluded from that family and suborder by their zygopteran bulging, hemispherical compound eyes set far apart on a short head (see Archibald et al. 2021). Wings most easily distinguished from those of other Zygoptera and from taxa that possess many of the nine wing character states listed below and might be Zygoptera or Cephalozygoptera as follows: from Viridiflumineagrion Nel by pterostigma ca. 2.5 times longer than wide [ca. 4.7 times], from Miopodagrion Kennedy by C-RA space distad pterostigma with one row of cells [apparently two full rows]. Wings distinguished from genera with similar wings without an associated head by one or more of the following: 10, C-RA space distad pterostigma almost always with a single row of cells, rarely a few rows two cells wide immediately distad pterostigma; 11, CuA-A space usually two, sometimes maximum three cells wide; 12, possession of a brace vein; 13, no accessory (“secondary”) antenodal crossveins:

Allenbya Archibald and Cannings : [10: 2 and 3 cells wide throughout];

Valerea Garrouste et al. : [10: many, up to five cells wide];

Thanetophilosina Nel et al. : [12: no brace vein];

Electrophenacolestes Nel and Arillo : [11: up to five wide, 13];

Menatagrion Nel and Jouault, 2022 [12, 13];

Chickaloon specimen of Garrouste & Nel (2019): [11: up to four wide];

Specimen MeI6572 ( Megapodagrionidae genus and species A of Petrulevičius et al. 2008, cf. Megapodagrionidae genus and species A of Archibald et al. 2021): [10: most of C-RA space two cells wide].

Emended description. Head short. Compound eyes bulging, hemispherical, set apart twice their width in dorsal aspect. Thorax, legs generalised as known by preservation. Forewing. Membrane darkly infuscate, but some might be hyaline as preserved (see below). Measurements, ratios of these see Tables 1 and 2 . Antenodal crossvein Ax0 absent, Ax1 and Ax2 somewhat lengthened to moderately expand antenodal space. No accessory antenodal crossveins. Pterostigma ca. 2–3 times longer than wide; anterior, posterior margins oblique; subtends 2–3 cells; distinctly oblique brace vein at basal-posterior corner in RA-RP1 space. Crossveins in postnodal, postsubnodal spaces mostly aligned basally, usually not distally. C-RA space distad pterostigma almost always with a single row of cells, rarely briefly two cells wide (see F. jutlandicus proposed neotype, Figs. 3–4 ); RA meets margin at or very near apex; slightly upturned near margin. Wing dense with cells throughout. RP2: origin 6.5–9 cells distal to origin of IR2. IR2: origin at or very near, basal to subnodus. RP3-4: origin about 3/5 from arculus to subnodus. Arculus at or very close basad Ax2. All major veins linear except MA zigzagged distad ca. mid-way between arculus, termination; CuA zigzagged, slightly near quadrangle, increasingly toward terminus. No crossvein O. CuA terminates on posterior margin ca. mid-way between nodus, apex; CuA–A space 2–3 cells wide at widest. Hind wing. Like forewing except shorter relative to width; measurements, ratios of these see Tables 1 and 2 . Abdomen generalised as known by preservation.

Remarks. Proposed neotype designation. Henriksen’s (1922) holotype and only specimen then known ( Fig. 1 ) is incomplete, consisting of two wings that are partially complete distal to the nodus, the mid-posterior fragment of a third, and an abdomen that is complete except for a portion of its base (MGUH 1819, Natural History Museum of Denmark, University of Copenhagen, Denmark). Lacking information from the diagnostically important antenodal region of the wings, he assigned the species to Phenacolestes , an extinct genus of the extinct suborder Cephalozygoptera known from the Eocene and possibly Miocene (see table 3 of Archibald et al. 2021) with similar venation in its preserved portions. Cockerell (1908) had assigned the genus to the Dysagrioninae, then a zygopteran subfamily of Agrionidae Leach (now Calopterygidae Sélys ). Recognising the limitations of the incomplete type specimen, Nel & Paicheler (1994) considered the species ‘ Phenacolestes ’ jutlandica as family indet. Rust (1999) expressed even less confidence in the generic assignment, treating the species as ‘Dysagrioninae gen. indet. jutlandica ’, illustrating it (his fig. 4 and plate 1, fig. a) with the non-type specimen FUM-N 13856 (then ERK KL Tl), an almost complete isolated wing and 14M-A2163 (his plate 1, fig. b).

Petrulevičius et al. (2008) assigned the species to their new, monotypic genus Furagrion based on ERK-KL-T1 (FUM-N 13856) ( Fig. 2 ). Zessin (2011) subsequently described Furagrion morsi and Morsagrion ansorgei , erroneously citing FUM-N 13856 (as ERK-KL-T1) of Petrulevičius et al. as the holotype of Furagrion jutlandicus . Archibald et al. (2021) considered Furagrion a dysagrionid in the Cephalozygoptera .

We agree that FUM-N 13856 is conspecific with Henriksen’s holotype by the extensive similarity of all preserved parts. Despite subsequently being treated as the specimen of reference, this fossil was not designated a neotype, however, and lacks the head, necessary for assessing family and suborder affinities (see below). Therefore, to bring clarity and nomenclatural stability to the Furagrion concept, we will be requesting in a forthcoming Case to the International Commission on Zoological Nomenclature that they designate as neotype of Phenacolestes jutlandicus specimen MM-10752, which is a well-preserved specimen in dorsal aspect, with all four wings and body almost complete, including the faint but distinct left compound eye (the right is indistinct), and parts of three legs ( Figs. 3–4 ).

Family assignment and the dysagrionoid grade. The wings of MM-10752 and the FUM-N 13856 wing are consistent with all nine wing character states used in part for the Dysagrionidae diagnosis (Archibald et al. 2021), and Henriksen’s MGUH 1819 possesses those character states for the characters that are observable. These are (see Fig. 5 , shown on the Dysagrion lakesii Scudder wing):

1- crossvein O absent;

2- arculus at or closely proximad Ax2;

3- quadrangle broad, distal side longer than proximal, posterior side longer than anterior, distal-posterior angle oblique, proximal-anterior angle usually about 90°;

4- nodus at least a quarter wing length, usually more;

5- AA, AP branch before joining CuP, AA briefly free distad petiole;

6- RP3-4 originates ca. one to usually two thirds the length from arculus to subnodus;

7- antesubnodal space without crossveins (note: Nel & Jouault (2022) mistakenly read this as antenodal space);

8- CuA–A space expanded in middle, at least two cells wide;

9- CuA long, ends on posterior margin at mid-wing or further

These character states define the Dysagrionidae in combination with head character states of the suborder Cephalozygoptera (diagnosis of Archibald et al. 2021): width across eyes about twice the length from the anterior margin of antefrons to the posterior of the occiput; compound eyes more or less adpressed to head capsule, convex laterally but not hemispherical, their posterolateral corners extended posteriorly to varying degrees, sometimes even acutely; the distance between compound eyes at the level of the centre of the ocelli is about the width of one eye or less, i.e., the head is not shortened and distinctly extended laterally with bulging, hemispherical compound eyes as in Zygoptera .

Archibald et al. missed that Rust (1999) had found that although the compound eyes are indistinctly preserved in Furagrion specimen 16-B3618, they are present and widely separated as in Zygoptera (“Von den grossen Komplexaugen sind nur undeutliche Reste überliefert. Sie liegen, wie für Zygopteren charakteristisch, weit voneinander getrennt an den Aussenseiten des Kopfes”, p. 19). He did not illustrate this specimen, and its whereabouts is not known to us.

We examined several specimens that conform to Rust’s observation and, therefore, treat the genus as a zygopteran. The compound eyes are clearly preserved in 14M-A2163 ( Fig. 6 ) and are faintly but confidently preserved in the proposed neotype MM-10752 ( Fig. 3 ). These have the typical zygopteran shape, hemispherical, widely set and bulging outward, and the head is short. The compound eyes are not preserved in MM-10750 ( Figs. 7–8 ), but the remaining head capsule is short and wide as in Zygoptera . Such a loss of eyes in fossils may happen, especially in Zygoptera , as they protrude and are more fragile, apparently more easily degraded than the robust head capsule, or they might simply break away from the head in death. Such missing compound eyes can also be clearly seen in multiple specimens of Lestes ceresti Nel & Papazian from the Oligocene of Céreste, France, cf. the holotype MNHN. F.R07445 (Archibald & Cannings 2021 fig. 1B and 1C) and PNRL 2019 and PNRL 2021 ( Nel & Jouault 2022, figs 11A and 12; the head of L. ceresti PNRL 2020 in their fig. 11B appears too poorly preserved to confidently evaluate) and see Chalcolestes tibetensis Xia et al. ( Xia et al. 2022, figs. 3A and 4) and Nel & Zheng (2021 fig. 2B).

Some other Zygoptera possess wing venation with many of the nine diagnostic character states of Dysagrionidae . For example, the extant Argia funcki Sélys ( Coenagrionidae, Coenagrionoidea ) only differs by character state 6, and species of Austroargiolestes Kennedy ( Argiolestidae Fraser , ‘Calopterygoidea’) by 5 and 6 (both Fig. 5 ). The wing of the Eocene Viridiflumineagrion aasei Nel ( Fig. 5 ) differs by character state 6 (assigned to “‘Megapodagrionidae’ sensu lato ”, which is highly polyphyletic, see Dijkstra et a. (2014), and so we treat it as family indet.). The extinct zygopteran Oligolestes Nel & Escuillé (family indet., see below) bears all character states but 1, 6, and 9 and Eodysagrion Rust et al. ( Rust et al. 2008, Fig. 9 ) (provisionally in the Thaumatoneuridae : Huang et al. 2017) differs only by character state 3. Miopodagrion possesses a zygopteran head, but much of its overlapping and variably preserved wings are difficult to separate and interpret with confidence ( Fig. 9 ). These wings can be established to possess character states 2, 3, 4, 8 and 9, but 1, 5, 6, and 7 appear uncertain or unknowable in its single fossil. We also consider it to be of unknown family. Treintamilun Petrulevičius ( Frenguelliidae ) ( Fig. 5 ) shares all nine character states except character state 1 (absence of crossvein O), which cannot be assessed by preservation; however, Petrulevičius & Nel (2003) report this crossvein in the other described frenguelliid genus, Frenguellia Petrulevičius & Nel. Frenguelliids bear a distinctive CuP, indicating that that family might not belong to either the Zygoptera or Cephalozygoptera , but could belong to an undescribed suborder ( Petrulevičius & Nel 2003, Petrulevičius 2017). Combinations of dysagrionid wing character states are then found widely across even distantly related odonates.

Genera belonging to the Dysagrionidae sensu Archibald et al. (2021) or possibly so were previously assigned to a variety of extant zygopteran families by wing venation, highlighting the generalisation of these character states. These include the “calopterygoid” families Thaumatoneuridae , Pseudolestidae , Megapodagrionidae (as then defined), Amphipterygidae , and “ Agrionidae ” (= Calopterygidae ) (e.g., Scudder 1878, Campion 1913, Tillyard & Fraser 1939, Fraser 1957, Carpenter 1992, Nel & Paicheler 1994, Bechly 1996, Rust 1999, Nel et al. 2005 a, 2005b; Nel & Arillo 2006, Rust et al. 2008, Garrouste & Nel 2015, Nel et al. 2016; Zheng et al. 2016a, 2016b, 2017, Huang et al. 2017).

There is thus a “dysagrionoid grade” of wing venation found among Zygoptera and Cephalozygoptera , and following Petrulevičius & Nel (2003) and Petrulevičius (2017) on the status of Frenguelliidae , then even outside of these suborders. These dysagrionoid wing character states might be symplesiomorphies shared by common ancestors which date back at least to the early Jurassic ( Kohli et al. 2021, Suvorov et al. 2022) with their stem taxa possessing wings that might look very much like those of e.g., the Frenguelliidae Petrulevičius & Nel (or Congqingia Zhang ?).

The family-level phylogeny of extant Zygoptera has undergone considerable development since the largely unresolved cladogram of Dijkstra et al. (2014) (see Bybee et al. 2021, Kohli et al. 2021, and Suvorov et al. 2022) and the superfamily Lestioidea Calvert appears to be well supported as sister group to the remaining extant Zygoptera . Within the latter, Platystictidae Kennedy are sister to the remaining non-lestoid Zygoptera ( Bybee et al. 2021, Kohli et al. 2021, Suvorov et al. 2022). Neither Lestoidea nor Platystictidae include taxa that possess the dysagrionoid wing venation (e.g., Garrison et al. 2010). The presence of dysagrionoid character states in zygopteran taxa might also represent homoplastic reversals as adaptations to similar selection pressures. A thorough study of the deep-time evolution of these wing characters based on basal zygopteran-cephalozygopteran phylogeny falls well outside the scope of this paper, but should be the focus of a future study.

The family and suborder designations of dysagrionoid grade taxa. The heads of Dysagrion Scudder , Phenacolestes , Petrolestes Cockerell , Congqingia , Okanagrion Archibald & Cannings , and Okanopteryx Archibald & Cannings , are known and consistent with the Dysagrionidae and Cephalozygoptera concepts (Archibald et al. 2021; Archibald & Cannings 2021) and so are confidently established in that family and suborder.

Archibald et al. (2021) further mention the presence of antenodal crossvein Ax0 as a potentially important character for identifying possible Cephalozygoptera fossils where the head is unknown. Although the wing base where Ax0 is found is often poorly preserved or absent in fossils, the vein is found in Dysagrionidae and Sieblosiidae where this region is well-preserved, and also found in the Whetwhetaksidae Archibald & Cannings , strengthening the notion that they are closely related (Archibald et al. 2021, Simonsen et al. 2022). Ax0 is absent or covered by sclerotization in Zygoptera ( Bechly 1996, Rehn 2003), except found in two Eocene species of Euphaeidae Jacobson & Bianchi (Archibald et al. 2012) and Burmadysagrion zhangi Zheng et al. ( Zheng et al. 2016a) . The wing base is well preserved in several of the fossils studied here (e.g., Figs 2 , 16-18 , 21 ), and although we find sclerotization in the region (e.g., Fig. 21 ), we do not find evidence for the presence of Ax0.

The family affinities of Furagrion and Viridiflumineagrion are unknown, as their dysagrionoid wing venation (poorly known in Miopodagrion ) alone is insufficient to establish this, and other relevant characters are little known beyond aspects of their heads that establish them as zygopterans.

The family affinity of Oligolestes Schmidt is also unknown ( Nel et al. 2005a; Nel et al. 2020; Nel & Zheng 2021; Nel & Jouault 2022). These authors compared it to the Sieblosiidae Handlirsch , but excluded that family as its diagnosis ( Nel et al. 2005a, p. 223) consists of a single character state that Oligolestes lacks: “highly specialised nodus apparently traversed by ScP, as the terminal kink of CP is shifted basally together with the nodal and subnodal veinlets and the nodal membrane sclerotisation is reduced”. Nel et al. (2005a, p. 223) concluded that “ Oligolestes could be closely related to the Sieblosiidae sensu stricto, but there is no proof supporting this hypothesis because all these characters are individually present in other damselfly lineages.” We agree with Nel & Jouault (2022) that Oligolestes stoeffelensis Nel et al. is a zygopteran by its distinctly zygopteran head and with all of the above authors that the genus cannot be assigned to a family given current knowledge. Further, the Cephalozygopteran head is clearly seen in the Sieblosiidae (see Stenolestes Scudder : Stenolestes cf. fischeri Nel MNHNF-B.47288 and less clearly in Stenolestes falloti (Théobald) holotype MNHN.F.B24507, Archibald & Cannings 2021, fig. 2a, b).

Odonates with dysagrionoid wings where the head is unknown are also then family and suborder indet., and at most might be considered cf. Cephalozygoptera , Dysagrionidae . These include Primorilestes Nel et al. ; Electrophenacolestes Nel & Arillo ; Stenodiafanus Archibald & Cannings ; Menatagrion Nel & Jouault; Allenbya Archibald & Cannings ; Thanetophilosina Nel et al. ; Valerea Garrouste et al. ; the unnamed Alaskan Chickaloon specimen of Garrouste & Nel (2019); specimen MeI6572 of the Senckenberg Museum, Frankfurt, Germany, treated as Megapodagrionidae genus and species A by Petrulevičius et al. (2008) and as cf. Dysagrionidae gen and sp. A by Archibald et al. (2021); specimen NHMUK I.9866/I.9718 of Nel & Fleck (2014). The Whetwhetaksidae is cf. Cephalozygoptera with more confidence by the presence of Ax0 (see above).

Principle component analysis and variation within the genus. PCA results are shown in Fig. 10 . The “ratios” plot ( Fig. 10A ) is based on the eight calculated ratios only, i.e., is based on wing shape. Twelve specimens are included. Because ratios are the only characters used in this case, the plot was based on the variance-covariance matrix. The variables are shown here as straight, green vectors spreading from the centre of the plot. Greater length of a vector (e.g., the RP2-apx/w ratio) shows more significant impact on the location of each specimen in the plot and the location of each specimen in relation the vectors show which ratios are most important in individual specimens. For example, FUM-N 14704 is in the lower right quadrant, indicating that the Arc-apx/RP2-apx and Arc-apx/w ratios are high (see also Table 4 ).

The “combined” plot ( Fig. 10B ) includes six size related data (measured distances of wing elements) and eight wing shape related data (ratios of these). The PCA calculations were based on the correlation coefficient matrix. Twelve measured wings were included, all with a full data set ( Table 4 ). The variance is almost equally distributed along the two shown axes with 40 % of the variance between the points occurring along the horizontal PCA-1 axis, and 38 % along the vertical PCA-2 axis. The results show that:

1- In the “ratios” plot almost all of the total variation (91%) is explained by the PCA-1, horizontal axis, while 6% is explained by the PC-2, vertical axis. In the plot comprising both ratios and direct measurements the values are 41% and 36% respectively.

2- MOA 770, the type of F. ansorgei , is located to the far right in both plots, supporting its separation from F. jutlandicus by its relatively narrower wing.

3- The more slender MOA 770 wing is probably a forewing, assuming the same forewing/hind wing differences as in F. jutlandicus .

4- MGUH 34113 is located very close to MOA 770 in the “ratios” plot ( Fig. 10A ), but not in the “combined” plot ( Fig. 10B ), showing that while the shapes of the two wings are very similar, their sizes are not.

5- The far-right position of MGUH 34113 and MOA 770 in Fig. 10A results from higher values of the ratios arc-pt/w and nod-pt/w than in F. jutlandicus ( Table 3 ).

6- The arc-apx/RP2-apx ratio ( Table 4 ) is larger in MGUH 34113 and FUM-N 14704 than in the remaining F. jutlandicus specimens.

7- FUM-N 11616 and FUM-N 13856 plot closer to known F. jutlandicus hindwings than forewings when only ratios are considered ( Fig. 10A ) and are characterised by arc-apx/w and arc-apc/RP2-apx values similar to known hindwings of F. jutlandicus .

8- Known hind wings of F. jutlandicus are placed in the upper left of the “ratios” plot and far to the left in the “combined” plot by their generally smaller values than forewings of the ratios that are divided by width, i.e., they are generally relatively wider than forewings.

Figures

FIGURE 1. MGUH 1819, Henriksen’s type specimen of Phenacolestes jutlandica: A, photograph; B, drawing; C–E, labelled wings. A, B to scale, C–E to scale, both scales 5 mm.

FIGURE 2. FUM-N 13856, with which Petrulevičius et al. (2008) defined Furagrion (as ERK-KL-T1): A, photograph; B, drawing. Both to scale, 5 mm. A forewing. Ax1, Ax2 = antenodal cross veins 1 and 2; arc = arculus; b = brace vein; CuA = anterior branch of the cubitus; IR1, IR2 = intercalar vein 1 and 2; MA, MP = the anterior and posterior branches of the media; nod = nodus; pt = pterostigma; Q = quadrangle; RA = radius anterior; RP1, RP2, RP3-4 = branches of the radius posterior; sn = subnodus; * = origin of RP2.

FIGURE 3. Furagrion jutlandicus proposed neotype MM-10752:A, photograph of the whole insect; B, photograph of the head; C, drawing of the head, showing bulging left compound eye. A to scale, 5 mm; B, C to scale, 5 mm.

FIGURE 4. Wings of Furagrion jutlandicus proposed neotype MM-10752. The top two are hind wings and the bottom two are forewings. All to scale, 5 mm.

FIGURE 5. The wing of the Eocene Dysagrion lakesii showing the nine character states of Dysagriondae (see text) and wings of other odonates bearing many of these. Orange indicates the antesubnodal space, blue the quadrangle, and green the CuA- A space. Asterix indicates the origin of RP3-4. All scale bars 5 mm, Argia funki (Selys) and Austroargiolestes brookhousei Theischinger & O’Farrell to same scale. D. lakesii redrawn from Scudder (1890), Viridiflumineagrion aasei from Nel (2022), Austroargiolestes brookhousei from Tann (2021), Argia funki from Garrison et al. (2010), and Treintamilun vuelvenlucha redrawn from Petrulevičius (2017). Abbreviations as in Fig. 2 except: AA = anterior anal vein; CuP = posterior cubitus; * = origin of RP3-4.

FIGURE 6. 14M-A2163: A, photograph; B, drawing; C. drawing of the hind wing (red in Fig. 6B). All to scale, 5 mm.

FIGURE 7. MM-10750: A, photograph; B, drawing, wing colours as in Fig. 8. Both to scale, 5 mm.

FIGURE 8. Wings of MM-10750, colour-coded forewings (FW) and hind wings (HW) to match colours in Fig. 7. All to scale, 5 mm.

FIGURE 9. Miopodagrion optimum (Cockerell), Priabonian of Florissant Formation, Florissant, Colorado, United States of America. A, full fossil; B, closeup of basal portion of the wings showing difficulty in confidently discerning, e.g., character states 5–7 of the dysagrionoid grade. Scale bars are 5 mm. Photo Talia Karim (University of Colorado, Boulder).

FIGURE 10. Principle component analysis of Furagrion wings. F. jutlandicus forewings in red, hindwings in blue, those unknown as fore- or hind wings in green, and the F. ansorgei wing is in purple. A, based on ratios only; B, based on both measured distances and ratios. See Table 4 for input data.

FIGURE 16. FUM-N 11696: A, drawing of bottom wing in 17C; B, drawing to top wing in 17C; C, photograph. All to scale, 5 mm.

FIGURE 17. FUM-N 14704: A, photograph; B, drawing. A forewing. Both to scale, 5 mm.

FIGURE 18. MM-2085: A, photograph; B, drawing. A forewing. Both to scale, 5 mm.

FIGURE 21. MM-11044: A, photograph; B, drawing.A fore- or hind wing. Both to scale, 5 mm.