Mavropanorpa, Willmann, 2024

Willmann, Rainer, 2024, Phylogeny and evolutionary history of Mavropanorpa n. gen. (Mecoptera: Panorpidae), Zoologischer Anzeiger 311, pp. 69-87 : 71-86

3. Characterisation of Mavropanorpa n. gen

Derivatio nominis: Mavros (gr.), black. Most body parts of the species belonging to Mavropanorpa are black in at least one of its morphs ( M. amurensis and allies are bivoltine with either black or light brown morphs).

Type species: Panorpa amurensis MacLachlan, 1872 (synonyms: P. orientalis MacLachlan, 1887 ; P. sachalinensis Matsumura, 1911 , Aulops melania Nav´as, 1912, P. jilinensis Zhou, 2000 and others).

Autapomorphic characters: Male: Hypovalvae short, emerging from a stalked common base (petiolus). Dististyli long as compared with Panorpa communis . The proximal sections of the subsidia („ventral parameres“ of authors) are thread-like and crossed (the left subsidium extends to the right side of the aedeagus and vice versa). Female: There is a tendency towards reduction of the apodeme of the medigynium. In M. amurensis and baohwashana it is still rather long ( Fig. 7 ), in M. japonica , lewisi and close relatives the apodeme has completely disappeared. Tendency towards reduction applies also to the lateral valvae. – Species of Mavropanorpa are large in overall size.

Species: M. amurensis (MacLachlan, 1872) ( type species; Sachalin, easternmost continental Russia, northeastern China, Korea), approximata (Esben-Petersen, 1915) ( Korea: Jeju; former European name: Quelpart [used by Issiki 1933: 407]; Saishiu in Okamoto 1925), babai (Miyamoto, 1979) ( Japan), baohwashana ( Cheng, 1957) ( China: Kiangsu [ Jiangsu]), chiensis ( Cheng, 1953) (Korea: Chi-i ), coreana ( Okamoto, 1925) ( Korea), japonica ( Thunberg, 1784) ( Japan: Honshu , Shikoku , Kyushu ), lewisi (MacLachlan, 1887) ( Japan), ishiharai ( Miyamoto, 1994) ( Japan: Shikoku ), kagamontana (Miyamoto, 1978) ( Japan), kellogi ( Cheng, 1957) (southeast China), kongosana ( Okamoto, 1925) (northeast China, Korea), miyakei (Miyamoto, 1979) ( Japan), nipponensis ( Navas ´, 1908) ( Japan: Honshu ), ochraceopennis (Miyak´e, 1910) ( Japan: Honshu , Shikoku , Kyushu ), okamotona ( Isshiki, 1927) (misspelled okamotoana by Issiki, 1933; north eastern China, Korea: Shakoji ), thompsoni ( Cheng, 1957) (= Panorpa tsushimaensis Miyamoto, 1979 ) ( Japan: Tsushima ), M. trizonata (Miyak´e, 1908) ( Japan: Honshu , Shikoku, Kyushu).

Genital structures of M. amurensis were described and illustrated using scanning electron microscope photos by Ma et al. (2010) under the name Panorpa jilinensis . Cheng (1957) gave a detailed description of M. japonica and figured male and female genitalia and the fore wing (Figs. 85, 90, 100, 104, 286). The species was said to occur in Japan and in China (Tien-tseun, Esben-Petersen, 1921: 43, Cheng, 1957: 47), but the Chinese record is an error.

Differences between Mavropanorpa and the type species of Panorpa , P. communis .

In both Mavropanorpa and Panorpa , wing colouration (if present) follows a common general scheme. There is an apical spot and a pterostigmal band which is often branched. However, in Panorpa communis the bands are narrower. In P. communis , the meso- and metanota have a characteristically shaped light band along the middle line, which does not occur in Mavropanorpa ( Fig. 1 ; exceptions are kongosana and okamotona where it is, however, a broad yellowish median band). In Mavropanorpa , the male abdominal segments 6–8 are much stronger developed than in P. communis , and sternite 9 has a hypovalval stalk ( Figs. 2 , 6a and 13 etc.). The dististyli of Mavropanorpa are longer in relation to the basistyli. P. communis has broad plate-like subsidia ( Fig. 6a ; in close relatives of communis they may be slender, strongly sclerotized braces, Willmann 1977) and they are not crossed. In Mavropanorpa the subsidia are narrow, almost string-like, and have a terminal enlargement. In P. communis , the aedeagal perticae („dorsal parameres“) bear a small transparent vexillum („flag“) that is more pronounced in Mavropanorpa . Prominent ventral valves on the genital field, which are characteristic of Mavropanorpa , are not developed in P. communis . The medigynium of female P. communis is large compared with that of Mavropanorpa ( Fig. 7 ), and in Mavropanorpa the apodeme as well as the lateral valvae of the receptacular papilla tend to get reduced. An apodeme may even be lacking ( Fig. 22 ).

4. Description of morphological genital structures with emphasis on Mavropanorpa

4.1. Male genital external structures

Male structures involved in copulation include abdominal parts that assist in bringing the two partners into the right position, tackle and grasp the female’ s abdominal tip und contribute to sperm transfer. Such structures are not only elements of the genital segment (the 9th abdominal segment and its appendages, the basi- and dististyli) but also the notal and postnotal organs on the 3rd and 4th abdominal segment, the cerci and even the wings. Wings are used as vibrating organs for intersexual communication and as devices that males lay over the females‘ wings so that they cannot fly away easily.

4.2. Segment 9

The 9th segment is split into the dorsal epiandrium (=epandrium; pre¨epiproct sensu Carpenter 1931; Cheng 1953, 1957; abdominal tergite 9) and the ventral hypandrium (sternite 9). Hypovalvae (hv, Figs. 2 and 3 ) are paired elongations of the hypandrium, usually covering much of the caudal genital foramen. The border between hypandrium and epiandrium is not visible.

4.3. Epiandrium

The epiandrium ( Fig. 8 ; epi) covers the anterior genital opening, the cranial genital foramen. In M. trizonata ( Fig. 8c ), ishiharai ( Miyamoto 1994 Fig. 10 ) and ochraceopennis ( Esben-Petersen 1921 Fig. 48) the epiandrium is terminally emarginated while it is acute in japonica and still more so in lewisi ( Fig. 8 a, b ) and nipponensis ( Esben-Petersen 1921 Fig. 47). In amurensis , coreana , approximata and baohwashana it is very long and bilobed, the lobes being relatively broad ( Fig. 8a–e ).

4.4. Hypandrium

The hypandrium splits into two caudally oriented hypovalvae in all Panorpidae . In Mavropanorpa the hypovalvae are not separate from the beginning of the elongation of sternite 9 as, for example, in Panorpa communis ( Fig. 6a ). Instead, sternite 9 begins as a unified elongation from which the hypovalvae arise only after a distance, thus forming a stalk (pt, petiolus; Fig. 2 ). Therefore, the hypovalvae are short as compared to the length of the basistyli.

A hypandrial petiolus evolved several times or/and was retained from remote ancestors independently in different species groups. Despite the latter statement the petiolus of Mavropanorpa is preliminarily considered to be one of the generic apomorphies as long as its sister taxon is not known with certainty because it might have a stalk as well (see below, remarks on P. jinhuaensis ). In M. amurensis and kongosana , it is broad or moderately broad ( Fig. 2 ; for M. kongosana , see Isshiki 1927 pl. 2 Fig. 5 and Issiki 1933 Fig. 7 E ). Presumably, this is the plesiomorphous character state relative to the japonica group (compare Fig. 23 ; M. lewisi , japonica , kellogi , trizonata etc.). Again, the hypovalvae of amurensis ( Fig. 2 ) and especially those of kongosana and okamotona are simple and almost parallel. In the japonica group they form more or less the splayed arms of a triangle. It appears plausible that they derive from forms such as those in okamotona and kongosana as do, on the other hand, those of the amurensis-approximata evolutionary branch. It appears that the hypovalvae of kongosana are near the plesiomorphous situation in Mavropanorpa . They are moderately long and almost parallel to each other, without any remarkable structures ( Issiki 1933: 328). The stalk is quite narrow. In okamotona the hypovalvae are much broader, and the stalk is wide.

In M. amurensis , the hypovalvae display a number of differentiations ( Figs. 2 and 3 ). The petiolus and/or the bases of the hypovalvae are broadened to form a medially directed triangular projection (see also Issiki 1933 Fig. 7G , Ma et al., 2010 Fig. 3 , Wang & Hua 2021 Fig. 6 F ). They end with a minute dorsally directed tip ( Fig. 3 ; not visible in ventral aspect). The latter structure is an autapomorphy of amurensis ; at least it was not described for any other species of Mavropanorpa . The series of strong, terminally bent setae along the inner margin of the hypovalvae is considered to be another autapomorphy. Such setae are recorded neither from M. kongosana nor okamotona ( Okamoto 1925; Isshiki 1927; Issiki 1933) nor any Japanese species of Mavropanorpa . M. approximata and baohwashana share tufts on the inner margin of the hypovalvae of ( Fig. 10a and b ).

In M. amurensis , the hypovalvae are broad and terminally rounded. In japonica and ishiharai , they are basally broad but distally they become narrower. In lewisi they appear to be intermediate. In Mavropanorpa lewisi and – according to Miyamoto 1994 Fig. 9 – M . ishiharai, both the petiolus and particularly the hypovalvae are broader than in trizonata , nipponensis and japonica .

The shape of the hypovalves is unique in baohwashana , chiensis and approximata ( Fig. 10 ).

4.5. Basistyli

The basistyli (bs) are fused and form a capsule („genital bulb“) with an anterior (cranial) and a posterior (caudal) opening, the „cranial“ and „caudal genital foramen“ ( Willmann 1981: 113; the terms derive from a proposal of Michener 1944). The caudal foramen displays the ventral external area of the aedeagus (gf, genital field [Genitalfeld, Grell, 1942]). The genital foramina are separated by the genital jugal bar (Genitaljugum, Willmann, 1981: 113), which connects the two basistyli on their dorsal side behind the cranial genital foramen. On each side, the basistyli have a dorsal medial process (prbs) behind the genital jugal bar ( Fig. 9a–c ; „interbasal plate of basistylus“ sensu Issiki 1933, „triangular process of gonocoxite“ sensu Wang and Hua 2021). A similar projection evolved independently in Panorpa ( Willmann 1981, 1989: 82–83; „Medialzapfen“).

In M. amurensis , japonica and kellogi the cleft between the basistylar medial processes is wide, but narrow and almost triangular in trizonata and lewisi in dorsal view. However, differences between the latter two species are remarkable. The attachment area of the epiandrial muscles terminates with a semi-oval form in trizonata . In contrast, the area is broad in lewisi and quite similar to trizonata and japonica . In amurensis it is of intermediate shape. The dorsal medial processes of the basistyli (prbs) are likewise different. They are simple protrusions in amurensis , japonica , kellogi and trizonata , but in trizonata and lewisi they are raised at least to the end of the basistyli ( Fig. 9b–c ). In amurensis and japonica ( Fig. 9a ) they do not reach the end of the basistyli. In lewisi , the processes differ from those in trizonata and the other species by forming a strong ridge (rprbs, Fig. 9 ).

The basistyli bear a pair of moveable appendages, the dististyli (ds). The basistyli are often called gonocoxites and the dististyli gonostyli, as it is likely that the two are modified abdominal limbs (basistyli + dististyli = gonopods). Snodgrass (1957: 35, 41), however and a few other writers substituted the terms basistyli and gonostyli by descriptive terms (basimere and telomere) because they considered the view that the external male genitalia were appendicular in origin as not tenable (also for Trichoptera, Lepidoptera and Diptera; contra Snodgrass 1935). The latter interpretation was not shared by most workers, including McAlpine (1981) for Diptera and Scudder (1971), who both reviewed the discussion.

4.6. Dististyli

The dististyli („harpagones“, „claspers“ etc. of some authors) of Mavropanorpa are long as compared to European species of Panorpa ( Figs. 6a and 13 , 17 , 18 ; for M. okamotona and M. kongosana see Isshiki 1927 Fig. 6 and Issiki 1933 Fig. 10 H , respectively). They are slender, but in M. amurensis they are, seen from the dorsal or ventral side, much broader than in the other species ( Fig. 2 ).

In M. japonica, kellogi and nipponensis , the distal inner margins of the dististyli are dentate (for nipponensis see Fig. 6 G in Wang & Hua 2021). In M. lewisi , M. ishiharai (figure in Miyamoto 1994), M. trizonata ( Willmann 1981 Fig. 13 c ), M. ochraceopennis ( Fig. 18 ) and M. okamotona ( Isshiki 1927 Fig. 6 ) there is a slight indications of such dentes. Plesiomorphic character state: inner margin without teeth ( M. amurensis , approximata , baohwashana ).

In amurensis , the dististyli are shorter than the basistyli ( Figs. 2 and 11 ). In M. ochraceopennis ( Fig. 18 ) and ishiharai ( Miyamoto 1994 Fig. 9 ) they are much longer. This is certainly derived compared with japonica , lewisi , trizonata ( Figs. 12 and 17 ; Willmann 1981 Fig. 13 c ) where they are as long as the basistyli. In kongosana and okamotona they are particularly long (apomorphy of the kongosana group).

Medituberculus of dististylus. In ventral or dorsal view the medituberculus of amurensis , baohwashana and kongosana is a broad triangular process that is set off from the basituberculus by a wide gap (mt, Fig. 2 ; for M. kongosana see Okamoto 1925 pl. 2 Fig. 5 and Issiki 1933 Fig. 10 H ; for M. baohwashana , see Cheng 1957 Figs. 82–83, Wang & Hua 2021 Fig. 9W ). (The situation in M. okamotona is not visible in the figure given by Isshiki 1927.) Comparison with species from various groups makes sure that this is the plesiomorphic state in relation to members of the japonica-group ( japonica , kellogi , ishiharai , nipponensis , ochraceopennis , trizonata , lewisi etc.), where the medituberculus lies close to the basituberculus. In nipponensis the medituberculus is especially close to it ( Wang & Hua 2021 Fig. 6 G ).

Basituberculus of dististylus. In lewisi and particularly in japonica , kellogi , ishiharai and trizonata the basituberculus is a slender, finger-like process (bt, Figs. 9 , 13 and 16 ). Plesiomorphic state: stout ( M. amurensis , Fig. 2 ; M. kongosana, Issiki 1933 Fig. 10 H ).

Mavropanorpa approximata , M. chiensis (both from Korea) and M. baohwashana ( China) have strongly and peculiarly curved dististyli. The hypovalvae begin as slender elongations of the hypandrium and are distally broadened. In chiensis , there is a basal small tubercle between the bases of the hypovalves ( Fig. 10 ). Doubtlessly, the three taxa are closely allied. M. a pproximata and baohwashana share a tuft of bristles on the medial side of the hypovalvae, and possibly a cleft in the stalk of the hypandrium (not shown for approximata by Esben-Petersen 1921, but see Issiki 1933 Fig. 7 F [see here Fig. 10 a ]).

4.7. Aedeagus

External structures for sperm transport into the females’ abdomen are the aedeagus and its parts.

The aedeagus (or, linguistically correct, aedoeagus) contains a sperm pump as described in detail by Grell (1942) and Willmann (1981). In Panorpidae the aedeagus is equipped with four external appendices which were often called „ventral“ and „dorsal parameres“ by mecopterists („titil[l]ators“ in the works of earlier authors; ventral and dorsal valves or „ventrovalvae“ and „dorsovalvae“ sensu Crampton 1918: 54, Carpenter 1931 and others, „ventral“ and „dorsal phallic valves“, Crampton, 1931: 11, „distal arms“ and „parameres“ in Snodgrass 1935: 600, „hypomeres“ and „epimeres“ in Ferris and Rees 1939 etc.). However, they have nothing to do with structures usually labelled parameres in Holometabola. These are the basi- and dististyli ( Crampton 1938, Nielsen 1957: 36, 53. The basi- and dististyli have been given a considerable number of names in other insects, e.g. Crampton 1922: 213–214, Snodgrass 1935: 599–606, Tuxen 1970. For a brief discussion, see above). Rottmar (1966) proposed the name „pseudoparameres“ for the aedeagal processes of panorpids, a term that may still leave the impression that they have something to do with parameres. Therefore, new names are introduced here: Aedeagal perticae (sg.: pertica, pertica aedoeagalis) for the dorsal processes, subsidia (sg.: subsidium, assisting element) for the ventral ones (prt and sub, Figs. 5a , 15 and 16a ). The two dorsal processes (one on each side) are terminal appendices of the aedeagus while the ventral ones, usually rooting in a lateral position, are only loosely attached to it. Shape of both the (dorsal) aedeagal perticae and the (ventral) aedeagal subsidia varies between species but it may do so to a certain degree within one and the same species as well. – For recent substitutions of the terms „dorsal“ and „ventral parameres“ in Diptera see Rice et al. (2019).

4.8. Aedeagal subsidium

Issiki (1933) notified that the males of several panorpids from east Asia are characterised by basally crossed subsidia (called ventral titilators by him), namely Mavropanorpa ochraceopennis , M. orientalis ( = amurensis ), and M. okamotona . Because of the crossing the left subsidium extends to the right side of the aedeagus, while the right subsidium extends to the left ( Issiki 1933: 335). This unusual character occurs, however, in several more species (see below). In Mavropanorpa , the subsidia are first (i.e. anteriorly) thread-like and immediately cross eachother. These are two characters that are very unusual within Panorpidae and without doubt a synapomorphy of all of its possessors (but see discussion of Panorpa jinhuaensis below).

In M. amurensis , the subsidia become very broad beyond the point of crossing (ventral aspect). Terminally, they are considerably enlarged, reaching the size of the aedeagal perticae ( Figs. 4 and 5 ).

In trizonata , the terminal portions of the subsidia are club-shaped, but in contrast to amurensis , they do by far not reach the size of the perticae ( Figs. 4 and 21 ). Again, they are much shorter than in amurensis .

In lewisi , the thread-like proximal section of the subsidium extends as far as to the terminal enlarged part ( Fig. 19 ). The latter is not club-shaped as in trizonata , but reminds of a very small blade that is apically pointed. This part leans on the aedeagal pertica reaching into a deep incision ( Fig. 19 ).

In M. japonica and kellogi , the aedeagal subsidia are different from all other species of Mavropanorpa . They are short sclerotized pointed hooks that nearly meet eachother almost in the middle of the aedeagal field ( Figs. 1 , 13–16a ). Because of their shortness, they do not cross. Their shape is certainly derived as other characters indicate a phylogenetic position amidst other species of Mavropanorpa . This implies that the lack of crossing in japonica and kellogi is secondary (autapomorphy).

The subsidia arise from latero-ventral areas of the aedeagus and are usually connected by a chitinous strut that extends into the basal part of the genital capsule ( Fig. 5b , described as „chitinous basal process of aedeagus“ in Issiki 1933, „Chitinbügel“ in German publications, „U-shaped bar“ by Snodgrass 1935: 600, „stalk of paramere“ in others). The subsidia are moveable by a pair of muscles running from the strut to the wall of the basistyli. Pending on the strength of the bond between the aedeagal subsidia and the central part of the aedeagus the subsidia can be moved more or less independently of the aedeagus proper. Both the ventral subsidia and dorsal aedeagal perticae are directed caudad, i.e. parallel to the basistyli.

4.9. Aedeagal perticae

The perticae emerge in a somewhat lateral position at the end of the aedeagus and are not movable by musculature of their own. Often, they have a pennant-like end (vexillum). In most species of Mavropanorpa , the vexilla of the aedeagal perticae are more or less club-shaped ( M. amurensis , ochraceopennis , trizonata etc., Figs. 5a and 18 ). In M. japonica , nipponensis , kellogi and lewisi (and quite certainly also in thompsoni and babai ) the perticae are deeply incised and therefore split into two branches ( Figs. 17 , 19 and 20 ; for nipponensis see Wang & Hua 2021 Fig. 10H ). Issiki (1933: 335, 406) interpreted the two terminal parts as independent processes of the aedeagus and called them „lateral“ and „dorsal titilator“ respectively. – The split is certainly derived as it does not occur in any other species of Mavropanorpa (and related scorpionflies). The terminal parts of the perticae are slightly different in the four species.

4.10. End of aedeagus and genital field

The ventral external side of the aedeagus has been called genital field. In Notiothauma ( Eomeropidae ) and probably primarily, it is composed of two parts, the genital folds. They meet in the middle of the genital organ but are not fused ( Mickoleit 1971). In Panorpidae , the folds are fused. The genital field bears the perticae and has on each side a lateral process (lp, processus lateralis aedoeagalis, Figs. 4 and 5a,b , 15 , 19 ) which articulates loosely with the inner walls of the basistyli. They function as gliding device and were, hence, described as such in Notiothauma by Mickoleit (1971: 343, „Gleitfl¨ache des Pumpenk¨orpersklerites“). Furthermore, the subsidia are attached to the ventral side of the aedeagus. Distally in the middle of the genital field lies the small opening of the ejaculatory duct, of which a sperm pump is a part. (Often the aedeagus is said to contain the ejaculatory duct. This would imply that the duct is but a hollow tube and has no tissue of itself, any tissue belongig to the wall of the aedeagus.)

The sperm pump consists of a pumping chamber, a moveable piston and its musculature. From the pumping chamber, sperm is pressed out by movement of the piston. The pumping chamber is constituted mainly of two strongly sclerotised parts and flexible connecting soft tissue that allows movement of the parts against eachother ( Fig. 16 ). The strongly sclerotised parts are (1) the end of the ejaculatory duct at the end of the aedeagal complex. This element forms the roof of the pumping chamber (rpch, Fig. 16b ) and bears differentiations on its external side (the genital field) as described above (e.g. the dorsal perticae). The piston (p, Fig. 16a ) is the second sclerite. It is part of the dorsal wall of the aedeagus and moveable by a pair of lateral promotor muscles and a medial depressor. In more plesiomorphous taxa, a further element can be distinguished, the piston carrier sclerite. In Panorpidae , it is almost indistinguishably fused with the piston. In Mavropanorpa , the pumping chamber is elongated in caudal direction to form a short tube behind the piston. This part of the genital chamber is the end of the ejaculatory duct where the genital opening lies.

In many panorpids, there are paired big bulbs in the proximal area of the genital field. They were called (aedeagal) ventral valves (vv, valvulae ventrales aedoeagalis, Figs. 4 , 5 , 15 and 19 ). They are not present in the type species of Panorpa and its close relatives. (Some authors, however, e.g. Crampton 1918; Carpenter 1931, used the term „valves“ for the structures here called aedeagal perticae [„dorsal parameres“ by many mecopterists].)

4.11. Abdominal segments 6-8

The relative sizes of male abdominal segments 6–8 have repeatedly been used to infer relationship. However, there may be considerable intraspecific variation, as specimens of Mavropanorpa amurensis from the continent and from Sachalin show ( Fig. 11 ). Therefore, in cases of minor differences a large number of specimens should be investigated before conclusions are drawn. For example, in M. okamotona , segments 6 and 7 are of almost equal length, while segment 8 is a little bit shorter ( Isshiki 1927 Fig. 6 ). The differences from amurensis from Sachalin is small. In M. amurensis and kongosana , segments 6–8 are longer, but almost of equal length ( Fig. 10 ; Okamoto 1925 pl. 5 Fig. 6 , Issiki 1933 Fig. 6 D ).

M. japonica has segment 6 elongated by almost 25 % as compared to M. amurensis . In M. lewisi , segment 6 is still longer. In lewisi , segment 7 is slightly longer than that of japonica ; therefore it appears to be less stout while it is actually of the same thickness as in japonica . Abdominal segment 8 is of the same length in both species. In japonica it is more slender than in lewisi ( Fig. 12a and b ).

4.12. Female genitalia

The main female genitals involved in copulation (vs genital structures involved in sperm storage, egg development etc.) lie in the 9th abdominal segment, in the genital chamber above the subgenital plate. The subgenital plate is the 8th sternite that shifted into the 9th segment ( Mickoleit 1975; 1978). The seminal duct (ductus receptaculi) ends in the membrane between segments 9 and 10. In Mecoptera , it is unique in having a complex sclerotized element, the medigynium (mdg, a term applied to panorpids at least since by Crampton 1929). Its main parts are a proximal apodeme (apo, Fig. 7 ; axis of some authors) supporting the seminal duct and a distal plate-like piece composed of several parts ( Mickoleit 1975; 1976). These parts include two terminal lamnae (la, Figs. 7 and 22 ) extending from either side (lateral or posterior processes of authors, valvulae sensu Grell 1942). Mickoleit assumed that the medigynium is a differentiation of the roof of the genital chamber but it also derives from the intersegmental membrane and the duct. A short apodeme is present in M. amurensis , kongosana and baohwashana (and quite certainly in okamotona , approximata and chiensis as well; females unknown) but it is completely reduced in the japonica group of Mavropanorpa ( Fig. 22 ). The lateral appendages of the papilla of the ductus receptaculi were labelled valvae of the receptacular papilla by Willmann (2022) in dependence on Mickoleit (1976), who had called them „flügelformige ¨Anh¨ange“ and justifiedly criticized earlier terms of medigynial structures. Another term was „lateral sclerite“ ( Mickoleit 1978). The medigynium of Mavropanorpa japonica was figured by Miyake´(1913 pl. 30 Fig. 9 ) (sub Panorpa klugi ). It is very similar to that of M. lewisi (here, Fig. 22 ).

Cerci of female Mecoptera are 3-segmented ( Fig. 22 ). The first cercal segment of the left and right side are medially fused („cercifer“ sensu Issiki 1933). Ventrally, a gap between the two basal segments is left. The laterotergites of abdominal segments 7 and 8 in Mavropanorpa females are longer than those in Panorpa communis , resembling those of Panorpodes , see figures in Mickoleit (1975).

5. Species groups and their characteristics

When Issiki (1933) described the subdistal crossing of the subsidia, he did not recognise it as a uniquely derived character shared by certain species. However, in a reconstruction of the phylogenetic relationships within Panorpidae by Wang and Hua (2021) thirteen species were included as members of a monophyletic entity based on this character (character no. 116 in Wang and Hua 2021. These are Panorpa kongosana , okamotona , ishiharai , japonica , nipponensis , ochraceopennis , amurensis , chiensis and baohwashana . For the remaining species that allegedly carry that character, see below). These species are said to share two more apomorphies: Abdominal segment 7 in the males is longer than segment 5 but no more than three times of segment 5 (character no. 46 in Wang & Hua 2021), and length of segment 8 (longer than segment 5 but no more than twice of its length (no. 54).

According to Wang and Hua (2021) this entity is comprised of three groups that were already proposed by Issiki (1933) plus the nikkoensis and waongkehzengi groups (for the nikkoensis group, see the remarks on P. takenouchii below; the waongkehzengi group contains P. waongkehzengi Nav´as, 1935, menqiuleii Wang et al., 2019 and jinhuaensis Wang et al., 2019 ). However, the relations of the nikkoensis and waongkehzengi groups remain unclear, and they may neither be monophyla nor be closely related with Mavropanorpa . Thus, the following three groups as identified by Issiki (1933) remain. (Species numbers differ from those given by Issiki and Wang and Hua because species that the three authors had omitted are also considered here).

- kongosana group: M. kongosana , M. okamotona . The kongosana group is characterized by dististyli that are much longer than the basistyli ( Issiki 1933: 407, no. 91- 1 in Wang & Hua 2021: 538)

- amurensis group: M. amurensis , M. coreana , M. approximata , M. chiensis , M. baohwashana .

- japonica group: M. ishiharai , M. japonica , M. nipponensis , M. ochraceopennis , M. kellogi , M. lewisi , M. babai , M. thompsoni , M. trizonata . For the japonica group, Wang & Hua (2021) listed two apomorphies: Abdominal segment 8 is more than twice as long than segment 5 (character state no. 54-2) and a series of small protuberances on the inner margin of the dististylus (no. 94-1). The latter character does not apply to all of the species, and in lewisi and japonica , segment 8 is as long as segment 5 or even slightly shorter ( Fig. 12a–b ).

The species included in the three groups are together a unit of most closely related species as shown by three apomorphies: a triangular process on the dorsal apex of the gonocoxites ( Wang & Hua 2021: character no. 89; prbs), a great swelling of the aedeagal ventral valves (126), and the shape of the aedeagal perticae (dorsal parameres of authors). The latter are described as „flattened and setose“ (138-1, Wang & Hua 2021: 540), a character description that is ambiguous.

The amurensis and japonica groups are said to share three characters. These are the shape of the notal organ that is slightly elongated but does not reach the middle of the 4th abdominal tergite (no. 27), the length of the 6th abdominal male segment (it is at least twice as long as segment 5; no. 37) and the setation of the inner margin of the hypopalvae (the setae are longer and stouter than those in other regions; no. 79). The first character is a plesiomorphy, not suitable for identifying natural entities. The described kind of setation of the hypovalvae does not occur in Japanese species of Mavropanorpa (see discussion of phylogenetic conclusions).

For Panorpa japonica, Wang and Hua (2021 : Fig. 15 ) viewed the character „ventral parameres [subsidia] not crossed“ as derived and characteristic of the species although the subsidia are almost never crossed in such a way in panorpids. This character state resulted from shortening of the subsidia to such a degree that they have become unable to cross and is a secondary character state. Therefore, Issiki 1933 wrote that „[i]n M. nipponensis and M. lewisi the aedeagus is similar to that of japonica . The only difference lies in the ventral titilators which are elongated and crossed so that the right titilator lies in the excavation of the lateral process of the left side, and the left lies in the right excavation of the lateral process of the right side“. Again, japonica is said to be characterized by only one A1-A2-crossvein in the forewing by Wang and Hua (2021, their character no. 19-2) but this applies to some individuals only ( Issiki 1933: 371) and can also be found in specimens of other species of Mavropanorpa including M. kellogi .

Remarks on P. takenouchii . Wang & Hua (2021) assumed that P. takenouchii , nikkoensis and dichotoma (all from Japan) belong to one of two evolutionary branches that are next to (though external of) the species groups that are here given the name Mavropanorpa . (The other branch is the waongkehzengi group.) Regarding takenouchii they viewed the crossing of the subsidia as homologous with that in Mavropanorpa . If homologous, this structure was decisive for assignment to Mavropanorpa . However, the subsidia of takenouchii are very different from those of Mavropanorpa : They consist of a big plate-like lateral appendage that is associated by a slim, strongly sclerotised medial structure. The slim medial structures cross not far from their bases (for figures, see Miyake´1912; Issiki 1933; Wang & Hua 2021). In contrast, the aedeagal subsidia of Mavropanorpa are single string-like sclerotised appendages that extend from lateral of the genital field (from the proximal chitinous strut towards the lateral process) and extend immediately after the point of its emergence towards the other side. The result is a basal crossing. – Other structures of takenouchii differ also from those in Mavropanorpa . The hypovalvae are far apart from each other, their stalk is short and broad (as in many scorpionflies), and the caudal genital foramen is extremely broad. The dististyli are much shorter than in Mavropanorpa . In the latter respect, takenouchii is plesiomorphous relative to Mavropanorpa . The medigynium (female) is also of unique shape. Therefore, without knowing where takenouchii belongs phylogenetically and, hence, systematically, the species is not included in Mavropanorpa . Its position and that of nikkoensis and dichotoma remains to be clarified.

Here, a new genus name is proposed with Panorpa takenouchii as type species.

6. Phylogenetic relationships within Mavropanorpa

For several structures under consideration, assumed character states (plesiomorphic vs apomorphic) were given above. The following lists apomorphies of species groups and species (compare Fig. 23 ). Close relationships between the species of Mavropanorpa may also be derived from the short subcosta in the forewing (it terminates far proximal of the pterostigma, meeting the costa) while a short vein, (Sc2) connects the costa and the first radial vein in the pterostigmal area. Such a course of the subcosta occurs also in Panorpa jinhuaensis , menqiuleii and waongkehzengi whose systematic position may be near Mavropanorpa . – The short subcosta led Enderlein (1910) and Navas´(1913) to assign species of Mavropanorpa (then Panorpa ) to Aulops ( typus generis: A. alpina Rambur, 1842 ; Europe), among them M. amurensis , japonica and nipponensis (sometimes under junior synonyms). Variation of the subcosta was dealt with e.g. by Esben-Petersen (1921: 13) and Issiki (1933: 353–358). Mavropanorpa corresponds to Issiki’ s section B of Panorpa species, and Issiki also recognized the kongosana-, amurensis- and japonica groups (see above). Furthermore, he distinguished two subgroups within the japonica group that are congruent with the results in the present paper.

Some characters in the following list of apomorphies, especially shape and course of the subsidia would get a different meaning if Panorpa jinhuanensis was the nearest relative of Mavropanorpa .

Monophyly of Mavropanorpa

- (1) aedeagal subsidia long and thin ( Figs. 5 and 10b, c , 19 , 21 )

- (2) subsidia crossed near their base ( Figs. 5 and 19 )

? - (3) end of subsidia enlarged, little sclerotized

- (4) basistyli with a medial dorsal process on either side (character no. 89- 1 in Wang and Hua 2021) ( Fig. 9 )

- (5) dististyli long, but shorter than basistyli ( Figs. 2 and 10 )

? – (6) hypovalvae with basal stalk (interpretation of character state pending on recognition of sister taxon of Mavropanorpa ) ( Figs. 2 , 13 , 17 and 18 )

M. amurensis + coreana + approximata evolutionary lineage (This is the amurensis-group sensu Issiki 1933: 407).

- (7) epiandrium with broad, incised end. This applies to M. amurensis , coreana , approximata and baohwashana . In chiensis , the epiandrium is almost acute, with a very small terminal incision; this shape is considered to be derived ( Fig. 8d and 10e ).

- (8) end of epiandrium bent ventrad at an almost right angle, reaching between the dististyli.

M. amurensis , autapomorphies

- (9) ventral side of dististyli with a row of spines ( Fig. 2 )

- (10) hypovalvae distally with small thorn-like process ( Fig. 3 )

- (11) medial sides of hypovalvae with a series of evenly distributed strong bristles ( Figs. 2 and 3 )

- (12) middle section of subsidia forming a broad band (ventral aspect) ( Fig. 4 )

M. approximata + baohwashana + chiensis (approximata lineage).

- (13) stalk of hypovalvae (petiolus) short ( Fig. 10 )

- (14) hypovalvae with slender basal portion, terminally enlarged, of unique shape

- (15) outer side of dististyli strongly bent mediad near base ( Wang & Hua 2021 character 93: 1) ( Fig. 10 )

approximata + baohwashana

- (16) stalk of hypandrium cleft along midline, resulting in very long hypovalvae. According to Issiki (1933: Fig. 7 F ) there is also an incision of the stalk in approximata (not shown by Esben-Petersen 1921: Fig. 57; Fig. 10a of the present paper contains both versions). - (17) medial sides of hypovalves with a tuft of bristles chiensis , autapomorphies:

- (18) hypovalvae with long slender basal section, enlarged terminal portion shorter than in approximata and baohwashana ( Fig. 10c )

- (19) short process of hypandrium between the bases of the hypovalvae ( Fig. 10c )

- (20) end of epiandrium narrow, with a very small distal incision ( Fig. 10d )

On Mavropanorpa coreana . M. coreana does not possess any of the apomorphies of amurensis and also not the peculiar shape of the dististyli of approximata , baowashana and chiensis . The short petiolus and the terminal enlargement of the hypovalvae (characters no. 13, 14) suggest a sister-group relationship with approximata + baowashana + chiensis .

Conflicting evidence in the approximata lineage and amurensis . In contrast to amurensis , the dististyli of approximata , chiensis and baohwashana are slender, thus resembling those in the kongosana and Japanese branchs of Mavropanorpa . Because of the peculiar shape of the epiandrium shared by amurensis , coreana and the approximata lineage (a supposed synapomorphy) slenderness of the dististyli is not taken as evidence of close relationship between the approximata branch and the kongosana and Japanese lineages. Either, slenderness is based on convergence, or it is a plesiomorphous heritage. The latter interpretation would imply that the broad dististyli of amurensis are derived. Initially, it appeared plausible that their broadness is plesiomorphic as a similar shape is common in scorpionflies outside Mavropanorpa .

The relatively stout subsidia of amurensis ( Fig. 4 ) are considered to be an autapomorphy, because slender, almost filiform subsidia occur both in the approximata lineage, which is supposed to be the sistertaxon of amurensis , and all other species of Mavropanorpa . If, however, filiform subsidia were derived as compared to those in amurensis , a different picture with amurensis as sistertaxon to the rest of Mavropanorpa would arise. Then, the terminally incised and ventrally bent epiandria of amurensis and the approximata lineage came out as a symplesiomorphy, and the stout shape of the dististyli of amurensis would also be plesiomorphic.

M. okamotona + kongosana + ishiharai + ochraceopennis + trizonata + japonica + thompsoni + kellogi + nipponensis + lewisi + babai .

- (21) dististyli slender ( Figs. 13 , 17 and 18 )

- (22) inner side of dististyli weakly dentate (more pronounced in japonica , kellogi and nipponensis , see below) ( Figs. 13 and 18 )

- (23) subsidia filiform almost over their entire lengths, terminating club-shaped ( Fig. 21 )

- (24) medituberculus of basistylus shifted towards basituberculus (further developed in the japonica group)

M. okamotona + kongosana

- (25) dististyli sickle-shaped, longer than basistyli ( Issiki 1933: 427, Wang & Hua 2021). This degree of length is considered to have evolved independently of ishiharai + ochraceopennis because synapomorphies shared by the subsequent species of Mavropanorpa indicate close relationship between the latter. Alternatively, the dististyli in other species of the japonica group could have become shorter.

M. ishiharai + ochraceopennis + trizonata + japonica + thompsoni + nipponensis + kellogi + lewisi + babai .

This is the japonica-group sensu Issiki 1933, who, however, could not consider ishiharai , kellogi and thompsoni because they were described after publication of his paper. He stated close relationship between japonica and nipponensis on the one hand and trizonata and ochraceopennis on the other. This is in agreement with the results in the present paper.

- (26) medigynium without apodeme (inferred from presence of apodeme e.g. in baohwashana and kongosana ; females of approximata and chiensis are unknown) ( Fig. 22 )

- (27) significant elongation of abdominal segment 8 (male) ( Fig. 12 ) - (28) stalk (petiolus) of hypovalvae narrow ( Issiki 1933: 406) ( Figs. 13 , 17 and 18 )

- (29) hypovalvae enclosing an open triangle ( Issiki 1933: 406: Hypandrium Y-shaped) ( Figs. 13 , 17 and 18 )

- (30) medituberculus of dististylus long, situated near basituberculus (plesiomorphic condition: a small triangular process, remote from basituberculus as e.g. in amurensis , kongosana , approximata and baohwashana ) ( Figs. 9 , 17 , 20 and 21 ; compare Figs. 2 and 10 )

- (31) basituberculus in dorsal or ventral aspect slim, almost fingershaped (plesiomorphous state as e.g. in amurensis , kongosana and approximata : stout)

M. ishiharai + ochraceopennis + trizonata .

? - (32) epiandrium (tergite 9) terminally emarginated ( Fig. 8c ). The character may not be apomorphous. An incision, though of another shape, is also present in amurensis , approximata and baohwashana ( Fig. 8d and 10e ) and still differently shaped in chiensis ( Fig. 10d ). If the emargination is not derived but a plesiomorphy, then a pointed epiandrium is an apomorphy of japonica + nipponensis + lewisi .

M. ishiharai + ochraceopennis

- (38) end of subsidia pointed. - Plesiomorphic character state: subsidia with terminal pennant (vexillum) ( Figs. 2 , 13–15 )

M. nipponensis , no autapomorphy identified. M. nipponensis is characterised by pointed subsidia that are still crossed basally (plesiomorphy relative to kellogi and japonica ).

M. kellogi + japonica .

- (39) subsidia short, not crossed ( Figs. 13–15 )

M. kellogi , autapomorphy:

- (40) stalk of hypovalvae very long, hypovalvae short

M. japonica , no autapomorphy identified.

Mavropanorpa thompsoni ( = tsushimaensis ) is closely related to japonica , but its exact position must remain unclear until the shape of its subsidia und perticae are described in detail.

M. lewisi + babai .

- (41) hypovalvae broad

- (42) ends of dististyli bent dorsad

- (43) wing markings reduced to large apical spot and minute spot immediately proximal of pterostigma ( Fig. 1e and f )

- (33) dististyli much longer than basistyli ( Fig. 18 )

M. trizonata , autapomorphy

- (34) wing markings consisting of large apical spot, broad pterostigmal band and broad submedian band ( Fig. 1 c-d). (Plesiomorphic state: submedian band small or consisting of single spots. Note that variation in M. japonica includes rare forms that may have similar wing markings.)

M. lewisi + babai + nipponensis + kellogi + japonica + thompsoni .

- (35) epiandrium terminally pointed ( Fig. 8a and b ). For interpretation of the character state, see above under M. trizonata + ishiharai + ochraceopennis and comments below

- (36) aedeagal perticae deeply incised, resulting in two branches ( Issiki 1933: 406: lateral titilators present) ( Figs. 14 , 15 , 19 and 20 )

M. nipponensis + kellogi + japonica + thompsoni .

- (37) inner margin of dististyli markedly dentate ( Fig. 13 )

Comments on the shape of epiandrium (tergite 9).

In amurensis , the epiandrium is broad and terminally deeply incised ( Fig. 8d ). It is similar to that in baohwashana and approximata . A terminal incision is also present in trizonata , ishiharai and ochraceopennis ( Fig. 8c ), but here it is a broad and shallow emargination, hardly comparable to that in amurensis . If (if!) the broad and shallow emargination is derived it would be an apomorphy of the three species. Interpretation is difficult as such a shape occurs in many species outside Mavropanorpa and might easily get lost or regained. However, a minute emargination was also described for chiensis ( Fig. 10 d ; Cheng 1953: pre-epiproct narrow towards apex, with very small U-shaped incision). In okamotona and M. kongosana ( Issiki 1933: 407) , there is no cleft (at the apex obtusely angulated, Isshiki, 1927: 6) as in japonica , kellogi , nipponensis and lewisi , where it is terminally almost pointed ( Fig. 8a and b ).

M. nipponensis and lewisi share an acute epiandrium. As it is suggested that this character can easily develop from less pointed epiandria, this (probably derived) feature is not taken as evidence for a sistergroup relationship between nipponensis and lewisi but assumed to be a convergently evolved feature. Characters indicating closest relationship between nipponensis and japonica are strongly dentate dististyli and acute subsidia.

6.1. Discussion of phylogenetic conclusions

The present reconstruction of phylogenetic relationships both supports and contradicts some of the results of others, e.g. of Issiki (1933) and Wang and Hua (2021).

While the study of Wang and Hua (2021) left the relationships among the Japanese species of Mavropanorpa unsolved, they are presented in this study. They are largely in agreement with what Issiki (1933) concluded. He distinguished two main groups, corresponding to the two main lineages identified here ( Fig. 23 , M. ishiharai – M. japonica ). In both the present work and the publication of Wang and Hua (2021) M. amurensis and the approximata lineage are found to be sister taxa (amurensis group sensu Issiki 1933), but the evidence is in part different. While Wang and Hua mentioned a shared subtriangular projection of the hypovalves (character no. 74-1), I am not sure about the homology of these processes which are, indeed, superficially similar. Furthermore, Wang and Hua relied on the degree of flexure of the distal half of the dististyli (which occurs also in approximata ). Such a flexion is also present in lewisi but their systematic distance makes it likely that the curves are a matter of convergence. A weak bend is also present in okamotona ( Isshiki 1927) .

Wang and Hua (2021) believed M. kongosana + okamotona to be sister of amurensis + the approximata lineage + the japonica group while according to the present paper it is sister to the Japanese branch only. Evidence for monophyly of amurensis + the approximata lineage + the japonica group (precondition for the sistergroup relationship supposed by Wang and Hua) is, however weak. The characters supposed by them to be synapomorphies are as follows: notal organ slightly elongated, not extending to middle of abdominal tergite 4 (their character no 27-1). However, a short notal organ not reaching the middle of tergite 4 is also present in okamotona + kongosana (which should be, according to Wang and Hua, outside the group with such a notal organ), and also in the waongkehzengi group. Again, in this size range, lengths (and general shapes) of notal organs differ gradually (even within species, Fig. 11 ; for Japanese Mavropanorpa , see Fig. 1 ). (37-2): abdominal segment 6 longer than segment 5 but not more than twice as long. This applies also to kongosana ( Issiki 1933: 325) . (79-1): setae along inner margin of hypovalvae longer and stouter than those in other regions. The character is difficult to conceive as Wang’ s and Hua’ s figures of alleged alternatives (2021: Fig. 6I and R ) shows hypovalvae with similar setae.

For M. miyakei , data sufficient for a phylogenetic analysis were not available.

7. Geography

Panorpa waongkehzengi , menqiuleei and jinhuaensis may be most closely related with Mavropanorpa , as Wang and Hua (2021) assumed, but they are not necessarily a monophyletic entity. Jinhuaensis appears to be, as the crossing of the subsidia shows, the sister taxon of Mavropanorpa , while the other two are more distantly related. The three species occur in eastern central China. The distribution area of jinhuaensis is closer to the eastcoast than those of the other two.

M. amurensis is widely distributed in continental northeast and on the island of Sachalin ( Esben-Petersen, 1921, Matsumura, 1911 and others). Despite some overlap in distribution areas, its nearest relatives within the genus are confined to more southern areas and consist of three main evolutionary branches. Members of the first evolutionary lineage are distributed in easternmost China ( baohwashana ) and Korea ( chiensis , approximata ). M. coreana appears to belong here as well. The second branch, consisting of kongosana and okamotona ( Fig. 23 ), occurs on the Korean peninsular and adjoining regions of China although it is not clear wether this branch is most closely related to the japonica group.

As said, Chinese M. baohwashana belongs to a branch that also occurs on the Korean peninsula with two species: chiensis and approximata (and perhaps coreana ). The distribution pattern is not difficult to explain as broad land connections between China and what is now the Korean peninsula occurred several times during the Pleistocene when sea level dropped repeatedly by more than 80 m during cold periods (about 120 m during the last glacial maximum). During one of these low sea level stands the volcanic island of Jeju must have been populated as well, where approximata developed in isolation ( approximata is confined to Jeju). Jeju came into existence about 2 mio years ago.

The time of origin of the kongosana-okamotona branch implies simultaneous genesis of the Japanese branch of Mavropanorpa , if it is its sister taxon. M. kellogi is the only known representative on the continent (southeast China: Fujian). The immediate ancestor of the Japanese lineage must have reached the archipelago coming from the south, either via the Korean peninsula or (which may be considered to be less likely) directly from China, also during a phase of low sea level via a land bridge that is now shallow sea.

Japan was probably populated by Mavropanorpa not earlier than during the first severe glaciations, as panorpids are poor flyers, unable to cross broad straits. Immigration by Mavropanorpa occurred only once and only by one species that became the stem of the Japanese branch as Japanese species of the genus are unified by several apomorphies. Again, as there are only minor morphological (and hence, genetical) differences between Japanese species of Mavropanorpa a quite recent time of origin is likely. Of course, other panorpids arrived independently of Mavropanorpa in Japan and possibly earlier during the complex geological origination of the Japanese archipelago since approximately 18 million years ago.

The differentiation of Mavropanorpa in Japan resulted in at least ten species (a few more species may have remained undetected), many of which can co-occur at one and the same several locality. For example, in the region of Sugadaira six species are present, plus additional five panorpids from other evolutionary lineages ( Suzuki and Ando 1987).

After separation of the two basal subgroups of the japonica group ( Fig. 23 ) and after nipponensis split off (that is, quite late in the history of the group) the ancestral species of japonica and kellogi developed hooklike subsidia that no longer cross. While japonica remained on the Japanese archipelago, the other daughter species reached the continent again; M. kellogi occurs in Fujian. It remains speculative which way ancestral kellogi may have taken. So far, no traces of the japonica group have been found elsewhere in China or on the Korean peninsula. However, another species closely related to japonica , kellogi and nipponensis , ancestral M. thompsoni populated the island of Tsushima. The exact phylogenetic position of M. thompsoni is not yet clear ( Fig. 23 ). Tsushima lies in the Korea Strait about 250 km north-east of Jeju. Arrival must have taken place after crossing what is now the Tsushima Strait (between Tsushima and Kyushu) during a quite recent glacial low sea level stand.

Figures

Fig. 7. Mavropanorpa and Panorpa, females, subgenital plate and medigynium, dorsal aspect. Left, M. amurensis, Gaolinzsy, ne China, 5-7-1949; right, P. communis, 3 km south of Nauen near Lutter, Germany, 2-7-2022. Middle: Medigynium of M. amurensis, same individual as in left fig., right lateral view. Scale 0.25 mm.

Fig. 1. Mavropanorpa n. gen., Japanese species. a-b M. nipponensis (Nav´as, 1908), a male, b female, Sugadaira, Ueda, Nagano pref., Japan; 9-7-2006. c-d M. trizonata (Miyak´e, 1908), Shinkazawa, Gunma pref., Japan; 9-8-1978. e-f M. lewisi (MacLachlan, 1887), Mt. Yokodake, Nagano pref., Japan; 10-8-1979. g-h M. japonica (Thunberg, 1784), Nagano pref., Japan.

Fig. 2. Mavropanorpa amurensis, male, abdominal sternite 9 and genital bulb, ventral view. Maoershan, ne China; 16-6-1950. Scale 0.5 mm.

Fig. 6. Panorpa communis (typus generis), male, Heiningen, Oker valley, Germany, 24-7-2022. a. Hypovalvae, basi- and dististyli and plate-like subsidia. Scale 0.5 mm. b. Same, male, 1 km west of Nienhagen, Mecklenburg-Vorpommern, Germany. Terminal abdominal segments, left lateral view for comparison with Mavropanorpa (Fig. 11-12).

Fig. 13. Mavropanorpa japonica, male, hypovalvae and genital organ, ventral aspect. The subsidia are secondarily shortened, see also Fig. 14–15.

Fig. 22. Mavropanorpa lewisi, female, terminal abdominal segments, ventral aspect. Subgenital plate removed to show medigynium. Left, medigynium from right side. An apodeme is not developed, compare with M. amurensis and Panorpa (Fig. 7). Scale 0.2 mm.

Fig. 3. Mavropanorpa amurensis, male, hypovalvae, left latero-ventral aspect. Scale 0.25 mm.

Fig. 8. Mavropanorpa, males, epiandria. a. M. lewisi, b. M. japonica, c. M. trizonata, d. M. amurensis. In amurensis, the apical portion (here shaded somewhat darker than the rest) is bent ventrad at an almost right angle. Scale 0.5 mm.

Fig. 10. Mavropanorpa, approximata evolutionary lineage, males. a–c: basi- and dististyli and hypandria with hypovalvae, ventral aspects. a M. approximata, b. M. baohwashana, c. M. chiensis. d–e: epiandria, dorsal view. d. M. chiensis, e. M. baohwashana. Not to scale. After Esben-Petersen, 1921, Issiki 1933, Cheng 1957a, Cheng 1953. As in Esben-Petersen’s monograph, fig. a does not show any aedeagal structures. Cleft of petiolus (stippled line) after Issiki 1933 Fig. 7 F. Figs c and d are each redrawn combining two figures of Cheng to show genital structures in one image per species; right hypovalvae are not depicted. Fig a shows the end of the epiandrium that is bent ventrad and thus visible between the dististyli. Figs d and e are less enlarged than a-c.

Fig. 5. a Mavropanorpa amurensis, male, latero-ventral view of caudal genital foramen and processes of aedeagus. Left lateral process shown translucent to show basal portions of left subsidium which terminates on the opposite side of the aedeagus. 5 b. Same, latero-ventral view of terminal portion of aedeagus, showing crossing of subsidia after removal from the genital capsule. Perticae of aedeagus not drawn (compare Fig. 5a). Scale 0.25 mm.

Fig. 23. Mavropanorpa, intrageneric phylogenetic relationships, most plausible solution. The japonica lineage is on the right (orange and black lines). Numbers are apomorphy numbers as in section „phylogenetic relationships within Mavropanorpa“. M. miyakei is unknown to the author and not included in the figure.

Fig. 9. Mavropanorpa spp., males, genital jugal bar and associated structures of basistyli, dorsal view. a. M. japonica, b. M. trizonata, c. M. lewisi. Scale 0.25 mm.

Fig. 17. Mavropanorpa lewisi, male, abdominal sternite 9 and gonopods, ventral view. Scale 0.5 mm.

Fig. 18. Mavropanorpa ochraceopennis, male, abdominal sternite 9 and genital bulb, ventral view. The drawing is based on a dried specimen. Nikko, Japan, 28- 7-1912. Coll. Esben-Petersen, Copenhagen. Scale 0.5 mm.

Fig. 11. Mavropanorpa amurensis, males, abdominal terminal segments, left lateral views. Above: Maoershan, China, below: Kusonrai, Sachalin, Russia, 2-8- 1925 (Esben-Petersen collection, Copenhagen). Note intraspecific length differences of segments 7–8.

Fig. 12. Mavropanorpa japonica (a) and M. lewisi (b), males, Sanada, Ueda, Nagano pref., and Mt. Yokodake, Nagano pref., Japan, 10-8-1979. Left lateral aspects of terminal abdominal segments, showing length differences.

Fig. 16. a Mavropanorpa japonica, aedeagus, right basistylus and basal part of right dististylus, caudal view; roof of pumping chamber removed. Scale 0.25 mm. b Roof of pumping chamber and end of aedeagus, cranial aspect. Scale 0.1 mm. The piston fits in the roof of the pumping chamber, at the end of which (in the dark shaded area) the opening of the sperm pump lies. Soft tissue on the sides of the chamber roof allows movement of piston relative to the chamber roof.

Fig. 15. Mavropanorpa japonica, terminal portion of right basistylus and basal portion of right dististylus, seen from left side, showing end of aedeagus and appendages after removal of left basi- and distylus. Arrow indicates level of cut between piston and roof of pumping chamber as depicted in Fig. 16. Scale 0.2 mm.

Fig. 4. Mavropanorpa amurensis, male, caudal genital foramen with genital field of aedeagus and its appendages, ventral aspect. Scale 0.25 mm.

Fig. 21. Mavropanorpa trizonata, male, caudal genital foramen, terminal portion of aedeagus with appendages and bases of dististyli, left latero-ventral view. Ends of subsidia club-shaped. Scale 0.25 mm.

Fig. 19. Mavropanorpa lewisi, male, caudal genital foramen and terminal portion of aedeagus with appendages, ventral aspect. The aedeagal perticae are branched. Scale 0.1 mm.

Fig. 20. Mavropanorpa lewisi, male, caudal genital foramen and terminal portion of aedeagus with appendages, left latero-ventral view. End of subsidia blade-like, extending into the cleft between the branches of the aedeagal perticae (only left pertica and left subsidium visible). Scale 0.2 mm.

Fig. 14. Mavropanorpa japonica, caudal genital foramen and terminal portion of aedeagus with appendages, left latero-ventral aspect. Scale 0.2 mm.