Silky lacewings (Neuroptera: Psychopsidae) from the Eocene-Paleocene transition of Denmark with a review of the fossil record and comments on phylogeny and zoogeography

STIG ANDERSEN

Insect Syst. Eval. Andersen, S.: Silky lacewings (Neuroptera: Psychopsidae) from the Eocene-Paleocene transi­ tion of Denmark with a review of the fossil record and comments on phylogeny and zoo­ geography. Insect Syst. Evo!. 32: 419-438. Copenhagen, December 2001. ISSN 1399-560X.

Thirty-four fossil insect wings from the Fur Formation of Denmark (Eocene-Paleocene tran­ sition) are recognised as neuropteran forewings standing close to extant Psychopsidae (silky lacewings). This affinity is based primarily on their possession of a vena triplica in the forewing. A new genus, PaZaeopsychops gen. n. and four new species, P latifasciatus sp. n. (type species), P abruptus sp. n., P angustifasciatus sp. n. and P maculatus sp. n. are described. PaZaeopsychops is identified mainly by the shape of vena triplica and costal space and by the venation of radial space. The new genus appears most closely related to some fos­ sil species of the North American Eocene, while it appears only remotely related to the psy­ chopsid genus Propsychopsis Kriiger described from Baltic amber (European Eocene). The different species of Palaeopsychops are separated on relatively minor differences in wing venation and pigmentation. The entire fossil record of psychopsid-like Neuroptera are reviewed and issues of phylogeny and zoogeography are discussed. A total of 5 extant and 21 fossil genera are recognised in the Psychopsidae lineage based on an extended vena triplica definition. The lineage goes back at least to the Upper Triassic. An attempt is made to gather the psychopsid genera into larger natural entities. Extant Psychopsidae are restricted to S. Africa, Australia and SE. Asia. It is suggested that the present relictual distribution is mainly caused by extinction events.

S. Andersen, Geological Museum, University of Copenhagen, 0ster Voldgade 5-7, D K -1350 Copenhagen, Denmark ([email protected]).

Introduction Oswald 1995). They are large, broad winged, noc­ The family name Psychopsidae is based on the turnal insects easily separated from other neur­ generic name Psychopsis proposed by E. Newman opteran families by the presence of a 'vena tripli­ in 1842. He derived the name from the Greek ca' formation in both the fore- and hindwings. 'psyche' meaning butterfly, and 'opsis' meaning Oswald (1993) describes vena triplica as 'the sc appearance (Oswald 1993). Indeed, the large size and r1 spaces run parallel, and are subequal in and the very broad wings with band patterns or width, from near the base of the wing to a com­ even eye-spots make psychopsids resemble but­ mon (or nearly common) distal point where each terflies (Fig. 1). However, the wings of psychop­ space is either constricted (by an interposed sids are very different in other features, e.g. the crossvein) or terminated (by brief vein fusion), transparent membrane and the dense network of and beyond (distal to) which point the parallel, longitudinal veins and costal veinlets bearing long subequal, nature of the spaces is not apparent'. macrotrichia on both surfaces and giving the The family is of special interest because of its wings a silky appearance ('silky lacewings'). relictual occurrence in Southern Africa, Australia, Psychopsidae are presently one of the smaller and Southeast Asia. families of Neuroptera comprising about 26 The first comprehensive treatment of fossil psy­ species in five genera (New 1989a; Oswald 1993; chops ids is that of Handlirsch (1906-08) who

© Insect Systematics & Evolution (Group 7) 420 Andersen, S. INSECT SYST. EVOL. 32:4 (2001)

Figures 1-2. (1) Adult male of Psychopsis elegans (Guerin), an extant species from Australia. (2) Detail of adult male of Propsychopsis sp. from Eocene Baltic amber. INSECT SYST. EVOL. 32:4 (2001) Silky lacewings from the Eocene-Paleocene Fur Formation 421 described several species from the Jurassic of flage when they were resting on tree trunks during Solnhofen and Mecklenburg in Germany. How­ daytime. Their larvae probably inhabited crevices ever, he referred them all to other families such as under bark where they were hunting other arthro­ Prohemerobiidae and Kalligrammatidae that he pods by their enormous, caliper-like mandibles. did not regard as closely related to extant The Danish psychopsid fossils are compression Psychopsidae. Tillyard (1922) was the first to fossils found as imprints (plate and counterplate) identify a fossil species as a psychopsid, viz. the in the so-called cement stone of the Fur Formation Triassic Triassopsychops superba from Queens­ in the western Limfjord of North Jutland. The Fur land, Australia. He noticed the presence of vena Formation (Pedersen & Surlyk 1983) is composed triplica in the forewing of Triassopsychops and of an approximately 60 m thick layer of diatomite, compared it with the extant Australian psychop­ the so-called 'mo-clay', which mainly consists of sid, Megapsychops illidgei Tillyard. Martynova diatomaceous shells and a small content of clay. (1949; 1962) has given the most comprehensive At certain levels, the diatomite is cemented with treatment of fossil psychopsids. Her systematic calcium carbonate and forming large calcareous conclusions have been followed by most subse­ concretions of cement stone. More or less con­ quent authors. spicuous, black layers of volcanic ashes further Fossil psychopsids have so far been recorded characterize the formation. The ash layers have from North America, Europe, Asia and Australia been numbered according to their relative ages and are probably also found in South Africa as from -39 (the oldest layer) to + 140 (the youngest indicated on the map (Fig. 13). The following 11 layer). The ash layers were deposited during the localities for fossil psychopsids are plotted on the Paleocene-Eocene transition and the age of the map: Ipswich, Queensland, Australia (Upper oldest layers has been estimated to about 54 Ma Triassic); Baisa, Transbaikalia, Siberia, Russia (,dinoflagellate zone 5', see Heilmann-Clausen (Lower Cretaceous); Krasnoyar district, Siberia, 1995). Certain ash-layers are so characteristic that Russia (Upper Cretaceous); Braunschweig and they can easily be traced from cliff to cliff and also Mecklenburg, N. Germany (Lower Jurassic); from the pits where the formation is exposed due Solnhofen, S. Germany (Upper Jurassic); Karatau, to commercial exploitation of the mo-clay. The Kazakhstan (Upper Jurassic); Fur and Mors, N. layers are best traceable in the cement stone. By Jutland, Denmark (EocenelPaleocene transition, far the largest source for insect fossils are loose about 54 Ma); Florissant, Colorado, USA (Upper rocks of cement stone which are found more or Eocene); Quilchena, British Columbia, Canada less abundant on the shores of the islands Fur and (Middle Eocene); 'samland' Peninsula, Russia and Mors and the adjacent mainland shores. Cement many other localities bordering on the Baltic and stones belonging to certain horizons are more eas­ Nqrth Seas (Baltic amber, Eocene, 50-40 Ma); ily partitioned than others and perhaps therefore Primor'ye Region, Russia (Lower Miocene). are apparently more fossil-rich. The psychopsid Most fossil specimens are preserved as com­ fossils are (when known) from the most fossil-rich pressions of wings (rarely other body-parts as cement stone horizon associated with the ash lay­ well) in sediments of various kinds. Only Propsy­ ers +25 to +30. chops is Kri.iger is known by three species (Mac­ The geology of the Fur and the correlated 01st Leod 1970) from extremely well preserved inclu­ Formation has been intensively explored for more sions of adults (Fig. 2) and larvae in Baltic amber. than 20 years, which has resulted in a good knowl­ They have been almost contemporary with the edge of the paleo-environment and the conditions new genus and species from the Danish Fur under which the fossils were deposited. In all Formation described below. probability (Bonde 1974; Larsson 1975; Andersen With a wingspan ranging from 7-9 cm the & Andersen 1996; Rust 1998; Andersen 1998; newly described fossil species from Denmark are Rust 1999) the land area where the insects lived among the largest psychopsids known. They are, was located near the southwestern coast of however, strongly exceeded in size by species of Norway from where they were carried out over the the Jurassic genus Kalligramma and allies with a sea by active flight or by the wind. Finally, they wingspan reaching 24 cm. The adults, as in extant were deposited along with marine sediments and species, were surely nocturnal. Their striped and vulcanous ashes in the epicontinental sea covering maculated forewings may have served as camou- all of Denmark in the early Paleogene. According 422 Andersen, S. INSECT SYST. EVOL. 32:4 (2001) to Rust (1999) the climate is supposed to have The fossils examined in the present work are been warm ('subtropical') and humid but very deposited in the Geological Museum. University likely also with dry periods during the summer. of Copenhagen (GMUC), Molermuseet = private During the 1990's a large and scientifically collections of Henrik Madsen and Bent S. important material of mo-clay fossils has been Mikkelsen, Skarrehage, Mors (MMSM) , and in found which has attracted several palaeontologists the Fur Museum, Fur (FMF). Very valuable mate­ interested in the Paleogene flora and fauna of rial is also kept in the private collection of Erwin Europe. The fauna comprises birds, turtles, Rettig, Nyk~bing Mors (ERNM). Two specimens snakes, starfish, snails, shellfish, about 50 species are kept in private collections in Germany. There of fishes, and more than 180 species of insects. A are no psychopsids in the smaller collections of detailed review of the diverse insect fauna is given insects from the Fur Formation in the Natural in an unpublished thesis by Rust (1999) while History Museum, London, U. K. (Andrew Ross, Schroder (1999) describes the nematocerous pers. comm.) and to my knowledge there are also Diptera in an unpublished report. The many new no material in the Palaeontological Institute, species and new genera described in these works University of Uppsala, Sweden. Thomas Schluter, will expectedly be published in a near future. The Nairobi, Kenya, borrowed in 1980 a large materi­ most recent survey of the palaeo-environment and al of Neuroptera also containing psychopsids from biodiversity of insects from the Fur and 01st the Fur Museum. Unfortunately, this material was Formations are provided by Andersen (1998) in a not available for study. monograph on fossil Gerromorpha (waterstriders). All line drawings are prepared from enlarged photographs and provided with details obtained Material and methods from careful study of all available material. All veins are drawn with almost equal thickness but as The first material of psychopsids from the Fur can be deduced from the photographs (Fig. 1-6) Formation came to my knowledge when my son the basal parts of veins Sc, R 1, Rs and Cu 1 are S~ren in 1994 found a well-preserved wing. The distinctly thicker than other veins. Macrotrichia, year after I found another wing in excellent condi­ microtrichia, trichosors and pigmentation pattern tion representing a different species. Apparently, are also omitted from the drawings. Some of the these two specimens represented the only existing photographs (Fig. 3-7) were taken from specimens material of psychopsids from the Fur Formation, wholly embedded in distilled water. as no specimens were identified as such in the col­ lections. It is curious therefore, that Larsson (1978: 121) stated that psychopsids are relatively Taxonomy common not only in Baltic amber but also in The PSYCHOPSIDAE lineage unsorted material of Neuroptera from the Fur (silky lacewings) Formation. Since my rediscovery of psychopsids in the Fur Formation a total of 31 wings (all The Psychopsidae lineage is presently defined as forewings) and 3 almost wholly preserved speci­ comprising the extant family Psychopsidae and its mens have been found in various collections. stem-group. The practical criterion for member­ Most of them were collected very recently but a ship of this lineage is apomorphy-based, viz. the few have been detected in old collections misiden­ presence of a vena triplica in the forewing. Details tified as wings of grasshoppers or cicadas. Some on this formation will follow. of the best preserved specimens came to my knowledge during my evaluation of insect fossils Genus under the Danish 'Danekrre' law (see Andersen & Palaeopsychops gen. n. Andersen 1996; Andersen 1998). This material of Etymology. - The generic name is derived from Latin psychopsids has been collected during the past 7- 'palaeo' = old and Psychops = the nominal genus of silky lacewings. 8 years by a few enthusiastic and skilled collec­ tors, notably Henrik Madsen and Bent S. Type species. - Pa/aeopsychops latifasciatus sp. n. Mikkelsen, the present and former heads of the Other included species. - Palaeopsychops abruptus sp. Molermuseum, Skarrehage, Mors, and a private n., P. angustifasciatus sp. n. and P. maculatus sp. n. collector Erwin Rettig, Nyk~bing Mors. INSECT SYST. EVOL. 32:4 (2001) Silky lacewings from the Eocene-Paleocene Fur Formation 423

Figures 3-4. Forewings of Palaeopsychops gen. n.: (3) P. latifasciatus sp. n. (Holotype). (4) P. abruptus sp. n. (Paratype No.3). 424 Andersen, S. INSECT SYST. EVOL. 32:4 (2001)

Figures 5-6. Forewings of PaZaeopsychops gen. n.: (5) P. angust~fasciatus sp. n. (Holotype). (6) P. macuZatus sp. n. (Holotype). INSECT SYST. EVOL. 32:4 (2001) Silky lacewings from the Eocene-Paleocene Fur Formation 425

Diagnosis. - Separated from other genera of the especially along apical and posterior margins lineage Psychopsidae by the following combina­ where only observable at high magnification. tion of forewing characters: Vena triplica tapered Costal space distinctly tapered towards apex of in distal third and terminating into a narrow back­ wing, with basal width (where broadest) about 3 wardly curved tip. Costal space basally about 2.5x times width of adjacent vena triplica and apical as wide as adjacent vena triplica and distinctly width (where narrowest) only about 2 times this tapered distally where at a point only half as broad width. Veins of costal space interconnected in one as basally and about 3x as wide as adjacent vena species by almost linear series of crossveins triplica. Vein M originating from vein R near basal (forming a gradate series, Fig. 10: grad.s); in furcation of veins Rl and Rs. Vein Rs with 27-31 remaining three species costal space without radial branches. Vena triplica with distinct crossveins; costal vein more or less curved, crossveins only present basally and with only two unbranched or branched up to four times. Veins crossveins in the subcostal space. Radial and (in Sc, RI and Rs distinctly thicker basally than sur­ one species) also costal space with crossveins rounding veins and running parallel except in dis­ forming gradate series. Differing from two unde­ tal third where gradually converging until merging scribed fossil taxa from Canada in the following of Sc and RI and finally merging of vein Sc+RI combination of characters (see discussion of rela­ and vein Rs to forming closed and narrow cell tionships below): forewing elongate, subtriangular with tip curved backwards (thus forming a distal­ in shape, 2.1-2.2x as long as wide (these traits are ly tapered and posteriorly curved vena triplica, see shared with one of the Canadian taxa); crossveins below); veins continue distally of vena triplica and fewer; individual branches strictly parallel and not branch 2-3 times before reaching wing margin. branched until near wing margin. Vena triplica with 2 crossveins between Sc and RI (Fig. 9: sc space) and 2-4 crossveins between Rl Description (Fig. 3-12). - Only preserved as iso­ and Rs (Fig. 9: r1 space). Spaces of r1 and rs of lated forewings except in three specimens with equal width. Rs giving off 27 -31 parallel branches superimposed wings and more or less well pre­ called radial branches and forming radial space served structures of head, thorax and legs (Fig. 8). (Fig. 9: r). Radial branches interconnected by two Long hairs of thorax preserved in two specimens. almost complete series of crossveins: one median Forewing: 2.7-4.5 cm in length (in recent psy­ and one distal gradate series (Fig. 9: grad.s). chopsids with lengths ranging from 1.0 to 3.5 cm), Individual radial branch branching again 0-3 elongate subtriangular in shape, 2.1-2.2x as long times between distal gradate series of cross veins as maximal width. Pigmentation pattern of wing and wingmargin. Vein M forked basally into veins unique among psychopsids: with numerous dis­ M 1 and M2. Veins Cu 1 (or CuA?) and Cu2 (or persed small spots and characteristic, more or less CuP?) stronger than surrounding veins and giving oblique, black bands, across the wing in three off each series of parallel branches distally; these species, restricted to large black spot and irregular branching again 2-3 times before reaching wing­ stripes in a fourth species. Hindwing shape and margin. Veins AI, M, CuI and Cu2 connected maculation pattern unknown, but in all recent basally by a few strong crossveins. Veins Al and species hind wing smaller and more evenly round­ Cu2 fused near furcation of Cu. Different appear­ ed than forewing and without a prominent macu­ ances of veins MI, M2 and CuI at their points of lation pattern or in few provided with relatively termination (near distal series of crossveins) pro­ small dark macula(e) distally. Membrane appar­ viding important specific characters. ently without microtrichia. All veins (except crossveins) provided with long dense hairs but Relationships. - Palaeopsychops gen. n. appears these only half as long as normally in recent psy­ most closely related to some fossil taxa of the chopsids and not preserved in most specimens North American Eocene, especially a species from (but usually represented by distinct scars as can be Florissant shales described as Polystoechotes seen in Fig. 7). In several specimens with distinct piperatus by Cockerell (1908). That species was outline of jugal lobe (as shown in Fig. 10). Wing subsequently assigned to Propsychopsis but sure­ margins with distinct trichosors (Fig. 7). Basal ly stands closer to the species here described in nygma between Rand M not detectable. PaZaeopsychops. Pterostigma absent. Venation extremely dense 426 Andersen, S. INSECT SYST. EVOL. 32:4 (2001)

Figures 7-8. (7) Basal part of forewing of Palaeopsychops abruptus sp. n. showing scars of hairs on veins and trich­ osors of costa (Paratype No.1). (8) Almost wholly preserved Palaeopsychops not referable to species because vena­ tion and pigmentation pattern on wing very faint and incomplete. INSECT SYST. EVOL. 32:4 (2001) Silky lacewings from the Eocene-Paleocene Fur Formation 427

~Ioo;:----ant.st ~~--eye

/-7--l.fr Hh-~c:r----H

~f---\:;:-~~-A 1 ~~~~-Cu2 I-+I-\-\--\,~~-Cu 1 IH+-T-\-r=-.::~-M 1 1++~,.-\--\o:;~~-M2 ~~-Sc ~~-R1 'I-+-\~~-Rs

-R-<'~-sC ~~-r1

~--cu----~~----r----~~c m v.tr Figure 9. Reconstruction of forewing of Palaeopsychops abruptus sp. n. showing very well preserved venation in a fragment (same specimen as in Fig. 7). Abbreviations of major veins and sectors of wing: Al - first anal vein, a - anal space, ant.st - antennal stumps, c - costal space, Cu - cubital vein, cu - cubital space, H - humeral vein, 1.fr - leg fragment, M - medial vein, m - medial space, Sc - subcostal vein, sc - subcostal space, r - radial space, Rl - first radial vein, Rs - branched radial vein, v.tr - vena triplica.

Palaeopsychops latifasciatus sp. n. cicada). Geological data: Cementstone, imprint situated very near a 5 cm thick ash-layer, probably ash-layer No. (Fig. 3, 10) +28. Deposited in FMF. Etymology. - Specific name derived from Latin 'latus' = Paratype No.2: Complete wing with well preserved broad and 'fasciatus' = banded. venation but less distinct pigmentation pattern on plate and counterplate. Length of wing: 3.9 cm. Locality data: Material. - Holotype: Complete wing with very distinct Fur. Collector unknown. Plate and counterplate marked venation and pigmentation pattern on plate and counter­ '12/3954 F2'. Geological data: Cementstone. Deposited plate (preservation exceptionally good on plate, Fig. 3). in FMF. Length of wing: 4.0 cm; about 2.1x as long as wide. Paratype No.3: Large fragment of wing with well Locality data: Gullerup Strand, Mors, Denmark. Date: preserved venation and pigmentation pattern on plate 12.III.1995. Erwin Rettig leg. Plate and counterplate and counterplate. Length of fragment 3.1 cm. Locality: marked 'Gu 12/3 95'. Geological data: Cementstone, Sydklinten, Thy. Henrik Madsen leg. Plate and counter­ probably ash-layer series Nos. +25 to +30. Deposited in plate marked '3-3560'. Geological data: Cementstone. ZMUC, but the fossil will be declared 'danekrre' and Deposited in MMSM. deposited in GMUC. Paratype No.4: Large fragment of wing with very Paratype No.1: Almost complete wing with very well well preserved venation and pigmentation pattern on preserved venation and pigmentation pattern. Length of plate and counterplate. Length of wing (estimated): 4.4 wing (estimated): 4.0 cm. Locality data: Fur. Magne cm. Locality: Fristurps Gl. Molergrav, Fur. M. Breiner Breiner leg. Plate marked' 1169', counterplate missing. leg. Plate and counterplate marked '204'. On enclosed On enclosed reposit card identified as 'sangcikade' (= reposite card identified as wing of '10vgneshoppe' (= 428 Andersen, S. INSECT SYST. EVOL. 32:4 (2001)

Figure 10. Reconstruction of forewing of Palaeopsychops latifasciatus sp. n. Abbreviations: grad.s - gradate series of crossveins, j.lob - jugal lobe. grasshopper). Geological data: Cementstone, ash layer tinuous across vena triplica (in closely resembling series Nos. +25 to +30. species P. abruptus the bands are more oblique Deposited in FMF. Paratype No.5: Almost complete wing with faint and discontinuous across vena triplica due to an impression of venation and pigmentation pattern on apparent displacement of costal sections). An plate and counterplate. Length of wing (estimated): 3.9 additional short fourth band present near apex of cm. Locality data: Skarrehage Molergrav, Mors. Henrik wing but this not well separated from black border Madsen leg. Plate and counterplate marked '11-A4498 , . Geological data: Cementstone, ash layer series Nos. +25 posteroapically. Black colour appears further as to +30. Deposited in MMSM. small spots and dashes dispersed all over wing Paratype No.6: Almost complete, superimposed surface, but most prominently on costal space wings and small fragment of thorax on plate and coun­ between bands. Costal space with gradate series of terplate (head absent); venation and pigmentation pat­ tern not well preserved and venation also disturbed by crossveins (in other species wholly without venation of underlying wing. Length of wing(s): 4.0 cm. crossveins); individual costal vein slightly curved. Locality: Ejerslev Molergrav, Mors. Henrik Madsen leg. Radial space with complete distal and less com­ Plate and counterplate marked' 14M-3029'. Geological plete basal gradate series of crossveins. Veins Ml data: Cementstone, ash layer series Nos. +25 to +30. Deposited in MMSM. and M2 slightly convergent distally but not com­ Paratype No.7: Fragment of wing with well pre­ bined before reaching the distal gradate series of served pigmentation pattern but rather indistinct vena­ crossveins (thus without a joint vein Ml +2). Vein tion on plate and counterplate. Length of fragment: 3.1 cm. Locality data: Sundby Klint, Mors. Henrik Madsen Cu 1 also not combined with vein M2 before leg. Plate and counterplate marked '5-4516'. Geological reaching distal gradate series of crossveins (thus data: Cementstone, ash layer series Nos. +25 to +30. without a joint vein M2+Cu 1). Deposited in MMSM. Paratype No.8: Almost complete wing with faint impression of venation and pigmentation pattern on Palaeopsychops abruptus sp. n. plate and counterplate. Length of wing: 3.7 cm. (Fig. 4, 9) Geological data: Cementstone. Private collection of Reinecke, Kiel, Germany. Etymology. - Specific name derived from Latin 'abrup­ tus' = disconnected. Description. - Forewing (Fig. 3, 10). Length: 3.7- 4.4 cm; about 2.2x as long as wide. Easily recog­ Material. - Holotype: Complete wing with very well nised by the following combination of characters: preserved venation and pigmentation pattern on plate and counterplate. Length of wing: 4.0 cm. Also contain­ Three relatively broad black bands run obliquely ing imprint of well preserved bug (Homoptera: across wing, each band almost straight and con- Pentatomidae) without wings. Locality data: Klitgarden, INSECT SYST. EVOL. 32:4 (2001) Silky lacewings from the Eocene-Paleocene Fur Formation 429

Mors. Date: 4.VIII.1994. Erwin Rettig leg. Plate and Material. - Holotype (Fig. 5): Almost complete wing counterplate marked 'KI 4/8 94'. Geological data: with very well preserved venation and pigmentation pat­ Cementstone, ash-layer series Nos. +25 to +30. At pres­ tern on plate and counterplate. Length of wing: 3.5 cm. ent deposited in ZMUC, but will be declared 'danekne' Locality data: Ejerslev Molergrav, Mors, Denmark. and deposited in GMUC. Henrik Madsen leg. Date: 21.1.1993. Plate and counter­ Paratype No.1: Plate with head, thorax, small frag­ plate marked '14M-4058'. Geological data: ments of leg and large fragment of two, partly superim­ Cementstone, ash layer series Nos. +25 to +30. posed forewings; head with distinct eye(s) and stumps Deposited in GMUC as 'danekne' No. 89 of antennae, forewings with extremely well preserved Paratypes No.1: Almost complete wing with very venation and pigmentation pattern (Fig. 7, 9); counter­ well preserved venation and pigmentation pattern on plate containing larger fragment but this very disturbed plate (less well-preserved on counterplate). Length of by underlying venation. Length of wing (estimated): 4.5 wing: 4.0 cm. Locality data: Ejerslev Molergrav, Mors. cm. Locality data: Mors. Plate and counterplate marked Mike Reich leg. Plate and counterplate marked 'ElE. '168-1998121'. Ole Burholt leg. Geological data: l.a'. Geological data: Cementstone, ash layer series Cementstone, ash-layer series Nos. +25 to +30. Nos. +25 to +30. Deposited in private collection of Deposited in GMUC as 'danekne' No. 168. Mike Reich, Greifswald, Germany. Paratype No.2: Almost complete wing with very well Paratype No.2: Almost complete wing with well pre­ preserved venation and pigmentation pattern. Length of served venation and pigmentation pattern on plate and wing: 3.7 cm. Locality data: Limfjorden. No other data. counterplate. Length of wing: 3.5 cm. Locality data: Plate marked 'Limfjorden', counterplate missing. Bygholm Vejlerne, Thy. Spren Andersen leg. Plate and Geological data: Cementstone, probably ash-layer series counterplate marked 'Psy.V.1994'. Geological data: Nos. +25 to +30. Deposited in GMUC. Cementstone, ash-layer series Nos. +25 to +30. Paratype No.3: Almost complete wing with well pre­ Deposited in GMUC. served venation and pigmentation pattern on plate and Paratype No.3: Almost complete wing with well pre­ counterplate (Fig. 4). Length of wing (estimated): 4.3 served venation and pigmentation pattern on plate and cm. Locality data: Bygholm Vejlerne (boulder), Thy. counterplate. Length of wing: 2.7 cm. Locality data: Stig Andersen leg. Plate and counterplate marked Klitgarden, Mors. Erwin Rettig leg. Plate and counter­ 'Psy.IY.1995'. Geological data: Cementstone, ash-layer plate marked 'KI T19'. Geological data: Cementstone, series Nos. +25 to +30. Deposited in GMUC. ash-layer series Nos. +25 to +30. Deposited in ERNM. Paratype No.4: Fragment of wing with rather well Paratype No.4: Large fragment of wing on plate and preserved pigmentation pattern and venation on plate counterplate with rather well preserved but faint impres­ and counterplate. Length of fragment: 3.0 cm. Locality sion of venation and pigmentation pattern. Length of data: Ejerslev Molergrav, Mors. Bent Spe Mikkelsen fragment: 3.0 cm. Locality data: ? Ejerslev Molergrav, leg. Plate and counterplate marked '1.863'. Geological Mors. Bent Spe Mikkelsen leg. Plate and counterplate data: Cementstone. Deposited in MMSM. marked 'I 684x'. Geological data: Cementstone, ash Paratype No.5: Small fragment of wing with rather layer series Nos. +25 to +30. Deposited in MMSM. well preserved pigmentation pattern and venation on Paratype No.5: Large fragment of wing on plate and plate and counterplate. Length of fragment: 3.0 cm. counterplate with rather well preserved but faint impres­ Locality data: Sydklinten, Mors. Henrik Madsen leg. sion of venation and pigmentation pattern. Length of Plate and counterplate marked '3-3560'. Geological fragment: 2.5 cm. Locality data: ? Ejerslev Molergrav, data: Cementstone, ash layer series Nos. +25 to +30. Mors. Bent Spe Mikkelsen leg. Plate and counterplate Deposited in MMSM. marked 'I 932'. Geological data: Cementstone. Deposited in MMSM. Description. - Forewing (Fig. 4, 9). Length: 3.7- Paratype No.6: Almost complete wing on plate and 4.5 cm; about 2.3x as long as wide. Wing closely counterplate with rather well preserved but faint impres­ resembling P. latifasciatus sp. n. and only distin­ sion of venation and pigmentation pattern. Length of wing: 3.5 cm. Locality data: Ejerslev Molergrav, Mors. guished from this species by the following unique Henrik Madsen leg. Plate and counterplate marked characters: 1) individual band distinctly discontin­ '14M-A3272'. Geological data: Cementstone, ash layer uous at level of vena triplica with costal section series Nos. +25 to +30. Deposited in MMSM. apparently disconnected (or displaced) from Paratype No.7: Almost complete wing on plate and counterplate with rather well preserved but faint impres­ remaining band; 2) costal space without any sion of venation and pigmentation pattern. Length of crossveins; 3) bands crossing wing more oblique­ wing (estimated): 4.3 cm. Locality data: Skarrehage ly; 4) veins Ml and M2 not or only slightly con­ Molergrav, Mars. Henrik Madsen leg. Plate and coun­ verged before reaching distal gradate series of terplate marked '11-A2053'. Geological data: Cementstone, ash layer series Nos. +25 to +30. crossveins. Deposited in MMSM. Paratype No.8: Large fragment of wing with well preserved venation and pigmentation pattern on plate Pa/aeopsychops angustifasciatus sp. n. and counterplate. Length of wing (estimated): 3.3 cm. (Fig. 5, 11) Locality data: Skarrehage Molergrav, Mors. Henrik Madsen leg. Plate and counterplate marked' 11-2776'. Etymology. - Specific name derived from Latin 'angus­ Geological data: Cementstone, ash layer series Nos. +25 tus' = nan·ow and 'fasciatus' = banded. to +30. Deposited in MMSM. 430 Andersen, S. INSECT SYST. EVOL. 32:4 (2001)

I M1 M2 Cu1 M1+2 Figure 11. Reconstruction of forewing of PaZaeopsychops angustifasciatus sp. n. Abbreviations: M 1+2 - fused MI and M2.

Paratype No.9: Large fragment of wing with well and counterplate with very distinct pigmentation pattern preserved venation and pigmentation pattem. Length of and venation but latter only distinct on pigmented area. wing (estimated): 3.3 cm. Locality data: Mors. J6rg Length of wing: 3.9 cm. Locality data: Ejerslev Ansorge leg. Plate marked 'MOA 742' (counterplate Molergrav, Mors, Denmark. Henrik Madsen leg. Date: missing). Geological data: Cementstone. Deposited in 3.VI.1996. Plate and counterplate marked '14M­ private collection of JOrg Ansorge, Dresden, Germany. B4885'. Geological data: Cementstone, ash layer series Paratype No. 10: Large fragment of wing on plate Nos. +25 to +30. At present deposited in ZMUC but will and counterplate with rather well preserved but faint be declared 'danekne' and deposited in GMUC. impression of venation and pigmentation pattern. Length of wing (estimated): 3.5 cm. Locality data: Description. - Forewing (Fig. 6, 12). Only known Silstrup, Mors. Henrik Madsen leg. Plate and counter­ from holotype. Length: 3.9 cm; about 2.2x as long plate marked '1-3816'. Geological data: Cements tone, as wide (the anterior margin is not visible as being ash layer series Nos. +25 to +30. Deposited in MMSM. partly embedded by the sediment). Pigmentation Description. - Forewing (Fig. 5, 11). Length: 2.7- pattern very different from other species as repre­ 4.0 cm; about 2.1x as long as wide. Easily recog­ sented by large black spot medially and several nised by the following combination of characters: irregular black stripes extending from spot to ante­ With four narrow bands of black pigmentation rior edge of wing; remaining surface of wing (bands only half as wide as in other banded apparently unpigmented. Veins of costal space species) and, in addition, a unique fifth narrow very curved (presence or absence of crossveins band crossing the other bands and following the could not be stated). Radial space with crossveins distal radial gradate series of crossveins. Costal forming two complete gradate series of space without crossveins; individual costal vein crossveins. Veins Ml and M2 not converged and slightly curved. Radial space with very long distal not fused before reaching distal gradate series of and rather incomplete basal gradate series of crossveins. Vein CuI combined with vein M2 in crossveins. Veins Ml and M2 converged and considerable distance from distal gradate series of fused into short joint vein Ml +2 before reaching crossveins and thus forming long joint vein distal gradate series of crossveins. Vein CuI not M2+Cul. combined with vein M2 before reaching distal gradate series of crossveins. Key to species of Palaeopsychops 1. Three to four isolated and parallel bands run­ Palaeopsychops maculatus sp. n. ning more or less obliquely from anterior to posterior edge of wing; without large medially (Fig. 6, 12) placed spot. Vein M2 and vein Cu 1 running Etymology. Specific name derived from Latin 'macu­ separately until reaching dista1 gradate series lata' = spotted. of crossveins ...... 2 - Six narrow, irregular, and short stripes extend­ Material. - Holotype (Fig. 6): Complete wing on plate ing anteriorly from large spot media1ly of INSECT SYST. EVOL. 32:4 (2001) Silky lacewings from the Eocene-Paleocene Fur Formation 431

Cu1 M1 M2 M2+Cu1 Figure 12. Reconstruction of forewing of Palaeopsychops maculatus sp. n. Abbreviations: M2+Cul - fused M2 and Cui.

wing. Vein M2 fused with vein Cu 1 and reach­ characteristics of the forewing include: '(1) costal ing distal gradate series of crossveins as joint space very broad throughout, in pterostigmal vein M2+Cu 1 provided with four parallel branches (Fig. 6, 12) ...... maculatus sp. n. region >2x as wide as the combined width of the 2. Four narrow bands and additional narrow, adjacent sc and rl spaces; (2) presence of only one crossing band, following crossveins of distal forewing nygma, located basally between veins R gradate series. Veins Ml and M2 fused shortly and M, (3) absence of pterostigmata, and (4) pres­ before reaching distal radial gradate series of crossveins and forming a short joint vein ence of at least five (usually> 10) crossveins tra­ Ml+2 (Fig. 5, 11) ...... angustifasciatus sp. n. versing the subcostal space (sc)' . - Three relatively broad bands not crossed by The current higher classification of extant narrow band apically. Veins Ml and M2 not Neuroptera places Psychopsidae in the suborder fused before reaching distal gradate series of cross veins ...... 3 Myrmeleontiformia that also contains the families 3. Bands crossing wing moderate obliquely; indi­ Nemopteridae, Myrmeleontidae, Ascalaphidae vidual band straight and continuous along and Nymphidae. This grouping is mainly based on length. Costal space with well developed gra- characters of the larval head as recognised by date series of crossveins (Fig. 3, 10) ...... latifasciatus sp. n. Withycombe (1924) and MacLeod (1964) but has - Bands crossing wing very obliquely; individ­ been corroborated by subsequent cladistic investi­ ual band distinctly discontinuous at vena trip­ gation (e.g., Henry 1978, Mansell 1992). Aspock lica with section in costal space appearing dis­ connected or displaced from remaining sec­ (1995) and Aspock et al. (2001) concluded that a tion. Costal space wholly without crossveins sister-group relationship exists between Psychop­ (Fig. 4, 9) ...... abruptus sp. n. sidae and Nemopteridae. The two families share the presence of a vena triplica in the hindwings. Unplaced specimens of Psychopsidae from the Fur Formation Classification of fossil Psychopsidae (mostly unidentifiable fragments) There exists little modem revisionary work on I 1081 (Mikkelsen); 11-2736 (Madsen); ll-A4498 fossil Psychopsidae. Most fossil genera currently (Madsen); 4-C3721 (Madsen); 5-C2127 (Madsen); 5- 4516 (Madsen); 1-187 (Mikkelsen; Fig. 8). referred to Psychopsidae (see New 1989b and Oswald 1993: 48) do not belong to the family in the narrow conception of Tjeder (1960), New Classification of extant Psychopsidae (1989a, 1989b) and Oswald (1993) based on the According to the diagnosis of the Psychopsidae extant fauna, but are assigned to Psychopsidae presented by Oswald (1993) all extant species of merely based on their possession of a vena tripli­ the family can be readily distinguished by the pos­ ca. According to Makarkin (1997) about twenty session of a vena triplica in the forewing. Other fossil species have been assigned to the Psychop- 432 Andersen, S. INSECT SYST. EVOL. 32:4 (2001)

sidae from Upper Triassic and up to the Miocene twice as wide as subcostal and radial areas put but most of them should for the present be placed together, or as wide; with strong cross vein sc-rl; similarity between costal area and anal margin of incertae sedis in Neuroptera. He only considered wing exists, but to a small degree; MP (= M2) many the following three genera Baisopsychops, Pro­ branched; posterior branch of MP forms a series of psychopsis and Triassopsychops as undoubtedly branches directed anteriorly; large number of 'among the Psychopsidae'. crossveins cover entire wing'. 3) Brongniartiellidae (as 'Brongniaritellidae'): 'Wings Handlirsch (1906-08: pI. 40, fig. 40-45, pI. 41, triangular; crossveins sc-rl and rl-rs absent in distal fig. 78-80, and pI. 48, fig. 1-3, 7) treated as Pro­ part of wing; CuA (= Cul) and CuP (= Cu2) with hemerobiidae several psychopsid-like species many branches; crossveins absent'. from the Upper Jurassic of Germany and England, some of which are certainly psychopsids. A few Bode (1953: fig. 321-343) placed several new others are possibly representing psychopsids but genera and species from the Upper Jurassic of are not well preserved (e.g. Archegetes and Crea­ Braunschweig in a subfamily Psychopsinae of groptera: Handlirsch 1906-08: pI. 48, fig. 1-3). Prohemerobiidae. Most of these taxa are here Martynova (1949) gave the first comprehensive excluded from the lineage Psychopsidae in the classification of fossil Neuroptera and also dis­ absence of a vena triplica. Only one of his species, cussed possible relationships of taxa with a psy­ Mesopsychopsis lias ina is certainly a psychopsid. copsid appearence. Based on their broad fore­ Another of Bode's species, the strange-looking wings and similar appearance of the veins SC Liassopsychops curvata (known only from large (Sc), R (Rl) and RS (Rs) (as being parallel and wing fragments) was considered closely related to equally spaced) she combined in a superfamily Triassopsychops and Apeirophlebia except dis­ 'Psychopsidoidea' the following four families: tinctly differing by characteristic S-shape (due to 'Psychopsididae', Kalligrammatidae, Brongniar­ distortion?) of the distal section of vena triplica. tiellidae and Osmylopsychopsidae. She provided a The species is in my opinion not sufficiently well short key to the families mainly based on subtle preserved to be assigned with confidence to the differences in the venation. The four families are psychopsid lineage. all synonymised below and assigned to the lineage SchWter (1986: fig. 3 and tab. 1) summarized Psychopsidae, a concept embracing the crown­ the results of Withycombe (1924), Martynova group + stem-group. (1962) and later authors in a phylogenetic tree Subsequently, Martynova (1962, English trans­ showing the stratigraphic range of fossil and lation 1991) no longer recognised the above four recent families of Neuroptera. He followed families as forming one superfamily but consid­ Martynova in placing Psychopsidae in the super­ ered them as members of a much larger superfam­ family Hemerobioidea and placing Kalligrammat­ ily 'Hemerobiidea' together with the families idae as the sister-group to Psychopsidae but did Palaemerobiidae, Sialidopseidae, Hemerobiidae, not mention any characters in support of this rela­ Sisyridae, Berothidae, Mesochrysopidae, Chrys­ tionship. SchWter did not consider Brongniartiel­ opidae, Prohemerobiidae and Proberotidae. She lidae and Osmylopsychopsidae as closely related provided the following short diagnoses for three to Psychopsidae. of the four 'psychopsidoid' families (extant and Carpenter (1992) followed Martynova (1962) fossil) found in Russia (therefore excluding in recognising the above-mentioned four families Osmylopsychopsidae from consideration): as members of a larger superfamily 'Hemerobi­ idea' , but he assigned the possession of a vena trip­ I) Psychopsidae: 'Wings large, triangular. In forewings costal areas three times wider than subcostal and lica solely to the family Psychopsidae. He added radial areas together; humeral vein short, bent the following new characteristics to Martynova's toward base of wing; distal parts of Sc, Rl and Rs diagnoses, most of which are here interpreted dif­ connected by crossveins which are as thick as the ferently or abandoned: 1) Psychopsidae: 'vena longitudinal veins they connect, distal connections thinnest; cubital-anal area forms a border along anal triplica present'; 2) Kalligrammatidae: 'Large in­ margin analogous to costal area; MA (= Ml) fused sects, related to the Psychopsidae', 'vein SC join­ with Rs; crossveins distributed at random along bor­ ing R near wing apex'; 3) Osmylopsychopsidae: der or form regular rows. Hindwing narrower than forewing'. 'as in Psychopsidae, but veins SC and R joined 2) Kalligrammatidae: 'Wing large (8-12 cm in length). well before apical region', 'crossveins few and Hindwings narrower than forewings; costal area weakly developed'; 4) Brongniartiellidae: 'as in INSECT SYST. EVOL. 32:4 (2001) Silky lacewings from the Eocene-Paleocene Fur Formation 433

Osmylopsychopsidae, but costal area narrowed about 4.5 cm long); the apparent forewing is tri­ markedly by vein SC terminating on C', 'distal angular in shape and shows a pigmentation pat­ branches of R and RS mostly straight, only a few tern, which is rather similar to the bands of curved posteriorly' . Palae()psychops. The two genera Sinopsychops (Hong 1982: fig. 134) and Beipiaopsychops (Hong 1983: fig. 86) Vena triplica as a synapomorphy for the from Middle Jurrassic of Beijang and Jiuquan of Psychopsidae lineage North China were both referred to the Psychops­ idae. However, as represented by small fragments The name 'vena triplica' was originally introduced of the wing only, they cannot be placed with cer­ by Tillyard (1918) to denote the 'strengthened' tainty in the psychopsid lineage. The same goes portion of the three 'parallel' main veins, Sc, Rl, and Rs in psychopsid fore- and hindwings. He for Epactinophlebia karabasica Martynova (1962: fig. 867) described in Brongniartiellidae; it is a also pointed out that the distal anastomosis of the small, psychopsid-like but poorly preserved fossil three veins in question is characteristic and helps from Upper Jurassic of Karatau, Kazakhstan. separating Psychopsidae (extant as well as extinct species) from other Neuroptera. Tillyard's obser­ vations were adopted by most subsequent authors Relationships of Palaeopsychops gen. n. (e.g. Tjeder 1960 and New 1989b). However, Os­ The four species presently described in the new ~ald (199~: 41) p~esented a somewhat deviating fossil genus Palaeopsychops on the basis of VIew on thIS venatlOnal formation: 'despite of the shared venational characteristics of the forewing fact that the possession of a vena triplica is an may belong to a monophyletic group (III, below) important part of the psychopsid wing gestalt it within the Psychopsidae lineage together with cannot be a psychopsid synapomorphy itself but is some contemporary fossil species from North best regarded as a visual perception created by the America. The first was described by Cockerell surrounding venation'. In his opinion it contains (1908: 59, pl.y, fig. 2) as Polystoechotes piberatus only two derived aspects: 1) the consistently broad in Hemerobiidae and is from the Upper Eocenel forewing and hindwing costal spaces, and 2) the Florissant scales of Colorado, USA. It was trans­ apparently abrupt termination, at the anastomosis, ferred to Propsychopsis KrOger in Psychopsidae of the linearity of the longitudinal veins, which by Carpenter (1943: 760) and was also considered com~rises the vena triplica. Oswald apparently as a psychopsid by MacLeod (1970: 166). Accord­ conSIdered the vena triplica of the psychopsid ing to Cockerell's brief description and photo­ wing as difficult to distinguish from the 'parallel graphic illustration, the supposed forewing is and robust' formation of the same veins in several about 3.0 cm long and only distinctly different other families of Neuroptera (e.g. many osmylids, from PaZaeopsychops in the pigmentation pattern. polystoechotids, ithonids, rapismatids and dilar­ A re-examination of the specimen and/or more ids). material is needed in order to clarify the status of . I disagree with Oswald (1993) that vena triplica this species. The other fossil species showing a IS not a proper character. However, the fossil evi­ good overall similarity with PaZaeopsychops are dence suggests to me that a wider definition of some undescribed fossils from the Middle Eocene vena triplica as originally intended is preferable: of British Columbia, Canada (Archibald 1998a, b) ~ena triplica is formed by the strengthened por­ represented by wings of two species: One consists tion of three equally spaced veins, Sc, Rl and Rs, presumably of a hindwing 4.5 cm long, as it is which are either parallel or converging distally without pigmentation pattern; it differs from and either terminated abruptly by an anastomosis PaZaeopsychops in having radial branches not of all three veins, or terminated more gradually by strictly parallel and provided with a larger number an attenuate, closed or narrowly open cell formed of crossveins. The other species is presumably by a varying degree of closure of Sc, Rl and Rs. represented by both fore- and hindwings (also This d:finition of vena triplica still defines an apo­ morphic state and allows the inclusion of some additional fossil taxa in the stem-group of 1 The Florissant Fonnation was previously considered Psychopsidae. of Lower Oligocene age but is here quoted as Upper Eocene following Andersen (1998: 8). 434 Andersen, S. INSECT SYST. EVOL. 32:4 (2001) Overview of taxa included in the aceous of Baisa, Transbaikalia, Siberia (Makarkin Psychopsidae lineage 1997: fig. 1-2). Based on one alleged synapomorphy, the presence I-d. Four genera from Upper Triassic of Ipswich, of a vena triplica formation in the forewing, a total Queensland, Australia: Archepsychops Tillyard of 5 extant and 21 fossil genera are here recog­ (1919a: 206, fig. 27), Triassopsychops Tillyard nised in the Psychopsidae lineage. They are provi­ (1922: 469, fig. 89; New 1989b: fig. 3), Osmylo­ sionally assigned to four groups (I-IV): psychops Tillyard (1923: 496, fig. 93) and Petro­ Group I. The first group is further subdivided into psychops Riek (1956: 104, fig. 3). four subgroups (a-d) that are surely not forming a The genus Osmylopsychops was originally de­ monophyletic entity, though they share a number scribed in Prohemerobiidae, while Martynova of characters: Wings short, subtriangular. Vena (1949), SchWter (1986) and Carpenter (1992) triplica with veins Sc, Rl and Rs almost parallel recognised it in its own family, Osmylopsychops­ from base to apex, where they either form a sud­ idae. den anastomosis or (in Triassopsychops) termi­ Group II. Kalligramma Walter comprising the nate in a short, backwardly curved tip. Costal largest known psychopsid, K. haeckeli Walter space very broad (>3x as wide as adjacent vena (Handlirsch 1906-08: pI. 48, fig. 12; New 1989b: triplica) and of equal width from base to apex. fig. 1, 2) from Upper Jurassic of Solnhofen, S. Costal and radial sectors with crossveins forming Germany, and several other species (Martynova costal and radial gradate series, except crossveins 1947: fig. 1; Panfilov 1968: fig. 1-5; Panfilov apparently absent in Osmylopsychops and 1980) from Upper Jurassic of Karatau, Kazakh­ Petropsychops. Rs with 10-21 parallel branches. stan. Meioneurites Handlirsch (1906-08: pI. 48, I-a. All extant genera of Psychopsidae: Psychopsis fig. 13; Panfilov 1968: fig. 6) from Upper Jurassic Newman (Australia, Fig. 1), Balmes Navas (SE. of Solnhofen, S. Germany; Kalligrammula Hand­ Asia), Silveira Navas (S. Africa), Cabralis Navas lirsch (1919: fig. 1; Panfilov 1968: fig. 8), Litho­ (S. Africa), and Zygophlebius Navas (S. Africa). gramma Panfilov (1968: fig. 7) and Kalligram­ The extant Psychopsidae constitude a well sup­ mina Panfilov (1980), all from Upper Jurassic of ported monophyletic group among the extant neur­ Karatau, Kazakhstan. Miopsychopsis Makarkin opterans (Oswald 1993). Even in terms of the (1991: fig. la-c) with two species from Lower known fossils the extant species are supported as Miocene of Primor'ye Region, Russia is tenta­ monophyletic by the characteristically abrupt dis­ tively included (known only from fragments of tal termination of vena triplica (absent from the wings with estimated lengths of only 3-4 cm but otherwise similar fossil genus Propsychopsis). with numerous crossveins especially in the radial space). I-b. Propsychopsis KrUger from Eocene Baltic The following combination of characters is Amber of NW. Europe (Kruger 1923; MacLeod found in this group of fossil taxa: Large species 1970: fig. 8-9, 12-14; New 1989b: fig. 14). with length of wing: 4.5-12.3 cm (except Miopsy­ Propsychopsis (Fig. 2) is known from adults of chopsis). Wings short and subtriangular (supposed three species and some undetermined larval spec­ forewing very broad and about as long as wide; imens, all extremely well preserved as embedded supposed hind wing narrower and about 2x as long in amber. These Eocene species show a close as wide). Veins Sc, R I and Rs parallel from base overall resemblance to the extant Psychopsidae, to apex except slightly converging distally. Vena especially species of the Oriental genus Balmes triplica partly open distally: Sc and Rl fused api­ according to MacLeod (1970: 170). Also the lar­ cally close to apex of wing and forming a back­ vae of Propsychopsis are very similar to the wardly curved tip, whereas Rs apparently runs known larvae of extant genera. However, the dis­ separately to the wing margin (yet connected to tal termination of vena triplica in Propsychopsis is Rl and Sc+Rl by several crossveins). Costal not abrupt as in all extant genera, but more like space of equal width from base to apex of wing, in that of Triassopsychops. Anyway, Propsychopsis (supposed) forewing about 1.6x and in (supposed) may stand closer to the extant genera than to the hindwing about 2.0x as wide as width of adjacent other known fossil genera. vena triplica. Venation forming a tight network of I-c. Baisopsychops Makarkin from Lower Cret- strictly parallel cross-veins between all larger INSECT SYST. EVOL. 32:4 (2001) Silky lacewings from the Eocene-Paleocene Fur Formation 435 veins (thus without gradate series of crossveins). lirsch: M. hospes Germar (Handlirsch 1906-08: pI. Rs with 5-12 parallel branches. 48, fig. 6) from Upper Jurassic of Solnhofen, S. The following characters are here considered as Germany, and M. lias ina Bode (1953: fig. 323) possible synapomorphies for group II: 1) costal from Lower Jurassic of Braunschweig in N. space narrow, not much broader than width of Germany; Apeirophlebia Handlirsch (1906-08: pI. adjacent vena triplica; 2) wing with dense network 41, fig. 82, only apex of wing) from Lower of crossveins; and 3) M2 (MP) consisting of sev­ Jurassic of Mecklenburg, N. Germany; Angar­ eral parallel branches. All the genera except opsychops Martynova (1949: fig. 10) from Upper Miopsychopsis were formerly placed in a separate Jurassic of Chita Province, Russia; Calopsychops family, the Kalligrammidae (better known as Panfilov (1980: fig. 95) and Propsychops Panfilov Kalligrammatidae) . (1980: fig. 96) from Upper Jurassic of Karatau, This group of fossil genera are known from Kazakhstan; Embaneura Zalessky (1953: fig. 1) Germany, Kazakhstan and Russia (Fig. 13). from Upper Cretaceous of W. Kazakhstan; Grammapsychops Martynova (1954: fig. 1) from Group III. Palaeopsychops gen. n. with three Upper Cretaceous of Siberia, Krasnoyar district, described species from EocenelPaleocene transi­ Russia. Only the genera Brongniartiella and tion of Denmark. 'Propsychopsis' piberatus Mesopsychopsis were formerly recognised in fam­ Cockerell from Upper Eocene of Florissant, ily Brongniartiellidae sensu Martynova (1962). Colorado, USA, a species evidently standing close The following combination of characters is to Palaeopsychops. Two undescribed species found in this group of fossil taxa: Wings short, (Archibald 1998a, b) from Middle Eocene of subtriangular. Vena triplica with veins Sc, Rl and Quilchena, British Columbia, Canada that may Rs gradually converging from base to backwardly represent distinct genera (see discussion above). curved apex and thus forming a long, narrow tri­ The following combination of characters is angle. Sc and Rl fused apically whereas Rs runs found in this group of fossil taxa: Wings elongate, separately to the wing margin. Costal space very subtriangular (about 2x as long as wide) except in broad basally (>3x as wide as adjacent vena tripli­ one of the species from Canada short, subtriangu­ ca) but strongly tapered distally of wing where at lar (only about 1.6x as long as wide). Vena tripli­ narrowest point only half as broad as basal width. ca with veins Sc, Rl and Rs strictly parallel in Crossveins absent in distal half of wing; basal 2/3 of length and in remaining apical 1/3 cross veins in basal half of wing few, not forming gradually converging and terminating into a nar­ typical costal and radial gradate series (except for row tip curved backwards. Sc and Rl fused short­ a few concentric series of crossveins basally in the ly before fusion of Sc+Rl with Rs, thus forming a wing of Grammopsychops). Rs with 16-26 radial fully closed vena triplica. Costal space basally branches, individual branch subsequently branch­ only about 2.5x as wide as adjacent vena triplica ed several times. but strongly tapered distally where at a point of The monophyly of group IV is supported by minimum only half as broad as basal width. Rs two characters: 1) vena triplica gradually tapered with 27-31 branches; individual branches parallel from base to apex; and 2) crossveins absent from and not branched until near wing margin except in distal half or more of wing. The distally tapered one of the Canadian species. costal space is shared with group III. The monophyly of group III is supported by: 1) The known distribution of group IV includes vena triplica tapered only in apical third; and 2) Rs Germany, Russia and Kazakhstan (Fig. 13). with as many as 27-31 branches (see Fig. 10). The group is distributed in Europe and North Taxa excluded from the Psychopsidae America (Fig. 13). As the distally tapered costal lineage space is a possible synapomorphy for group III and the Eurasian group IV considered next, an The fossil genera Actinophlebia Handlirsch, Eocene dispersal of the present group from Paractinophlebia Handlirsch and Protopsychopsis Europe to North America seems plausible. Tillyard have been placed in the Psychopsidae (Martynova 1949, 1962; Carpenter 1992; Ansorge Group IV. Brongniartiella Meunier (Handlirsch 1996). In the absence of a vena triplica in the 1906-08: pl. 48, fig. 5); Mesopsychopsis Hand- forewing, these genera are presently transferred to 436 Andersen, S. INSECT SYST. EVOL. 32:4 (2001)

Figure 13. World distribution of extant (shaded areas) and fossil (asterisks) Psychopsidae. Abbreviations: M ~ Miocene, E - Eocene, P ~ Eocene-Paleocene transition, C - Cretaceous, J - Jurassic and T - Triassic (South African fossil record according to Schluter in litt.). the extinct family Prohemerobiidae. The wings in triplica, attributed only a total of 11 fossil genera Prohemerobiidae belong to the most primitive to the Psychopsidae, but he noted that the fossil type in Neuroptera according to Tillyard (1919b). families Kalligrammatidae, Brongniartiellidae Apart from the absence of a vena triplica, prohe­ and Osmylopsychopsidae 'contain a number of merobiids also differ from members of the psy­ additional taxa thought to be closely related to chopsid lineage by the narrower sc space com­ psychopsids'. Based on his examination of the pared with the r 1 space. fossil genera he concluded: 1) that 'silky lace­ wings were morphologically diverse as well as geographically widespread during the Mesozoic', Phylogenetic conclusions 2) that 'they had their principal radiation possibly The presence of a vena triplica (in the broad con­ during the Jurassic', 3) that 'the oldest fossils ception of the present paper!) in the forewing is attributed to the Psychopsidae (Upper Triassic of the only derived character state that can be used to Australia) date the origin of the family prior to the identify compression fossils as members of the division of Pangea into Laurasia and Gondwana­ Psychopsidae lineage (crown-group + stem­ land', and 4) 'the presence of psychopsids on group). An Upper Triassic minimum age for the Laurasian land areas (North America: Colorado, lineage is confirmed on the basis of Triassopsy­ British Columbia; Europe: Baltic Amber) during chops and allied genera from Australia. The extant the Tertiary suggests that the disappearance of members of Psychopsidae share one supposedly silky lacewings from much of or most of ancient apomorphic character of the wing, viz. the abrupt Laurasia is a relatively recent event'. My assign­ termination of vena triplica (see wings of Psy­ ment of a considerably higher number of fossil chopsis in Fig. 1). The absence of this character genera to the Psychopsidae lineage tends to sup­ state in all known fossil psychopsids suggests that port Oswald's zoogeographic conclusions. A none of them belongs to the crown-group of Pangean origin for the Psychopsidae lineage is Psychopsidae. supported by the record of psychopsid genera from the Upper Triassic of Australia and assumed existence of psychopsids from the Triassic of the Zoogeographic conclusions Moltena Formation of South Africa (Schlitter in Oswald (1993), using a narrow conception of vena litt.; Fig. 11). The most significant new observa- INSECT SYST. EVOL. 32:4 (2001) Silky lacewings from the Eocene-Paleocene Fur Formation 437 tion is that Eocene Palaeopsychops from the Fur chena, British Columbia. Unpublished report, Formation relates more closely to Eocene psy­ Department of Biological Sciences, Simon Fraser University, Surnaby, B.C., Canada. chopsids of North America than to the almost con­ Archibald, B. (1998b). Fossil fly images on Internet temporary Propsychopsis from the Baltic amber. (also including other insects). Http://www.mjt­ The tripartite distribution (Australia - SE. Asia books.comlimages/neuroptl.jpg . - S. Africa) of extant psychopsids was explained Aspock, U. (1995) Neue Hypothesen zum System der Neuropterida. Mitteilungen der Deutschen Gesell­ by Oswald (1993: fig. 61-62) from an Australian schaft fUr Allgemeine und Angewandte Entomologie dispersal hypothesis saying that SE. Asian psy­ 10: 633-636. chopsids originated from Australia via dispersal Aspock, U., Plant, J. D. & Nemeschkal, H. L. (2001) through the Malay Archipelago. However, fossil Cladistic analysis of Neuroptera and their systematic position within Neuropterida (Insecta: Holometabola: evidence, especially the strong morphological Neuropterida: Neuroptera). Systematic Entomology similarity between extant Psychopsidae and 26: 73-86. Propsychopsis from Baltic Eocene, suggests that Bode, A. (1953) Die Insektenfauna des Ostniedersach­ extinction events have had a considerable impact sischen Oberen Lias. Paiaeontographica 103 (A): 1- 375. on the distribution of modern psychopsids. Bonde, N. (1974). Palaeoenvironment as indicated by the "Mo-clay formation" (Lowermost Eocene of Denmark). Tertiary Times 2: 29-36. Acknowledgements Carpenter, F. M. (1943) Osmylidae of the Florissant I thank Nils M. Andersen (University of Copenhagen) shales. Colorado. American Journal of Science 241: and Rainer Willmann (University of Gottingen) for 753-760. valuable comments on earlier drafts of the manuscript. I Carpenter, F. M. (1992) Arthropoda. Superclass Hexa­ am also very grateful to Henrik Madsen and Bent S0e poda. Treatise on Invertebrate Paleontology, Part R, Mikkelsen (Molermuseet, Mors) and a private collector Arthopoda 4. Vol. 2: ii + 279-655 pp. Geological Erwin Rettig (Mors) for permission to study their col­ Society of America & The University of Kansas. lections. Without their goodwill this work would not Cockerell, T.D.A. 1908. Fossil Insects from Florissant, have been possible. I also thank Jorg Ansorge Colorado. Bulletin of the American Museum of (University of Greifswald), Georg Stenstrop (Fur Natural History 24: 59-70, incl. pI. V. Museum) and Niels Hald (University of Copenhagen) Handlirsch, A. (1906-1908) Die Fossilen Insekten und for permission to study material under their care, and Jes die Phylogenie der Recenten Formen. 1430 pp. W. Rust and Florian SchrOder (both Gottingen University) Engelmann, Leipzig [1906: pp. i-vi, 1-640, pI. 1-36; for their interest and support. 1907: pp. 641-1120, pI. 37-51; 1908: pp. 1121-1425]. Nils M. Andersen and Morten W. Holmgren Handlirsch, A. (1919) Eine neue Kalligrammide (Copenhagen) helped me with computer graphics. The (Neuroptera) aus dem Solnhofener Plattenkalke. photographs were kindly provided by Sten Lennart Senckenbergiana 1: 61-63. Jacobsen (Geological Museum, University of Henry, C. S. (1978) An unusual ascalaphid larva Copenhagen) (Fig. 1), Christel and Hans Werner (Neuroptera: Ascalaphidae) from southern Africa, Hoffeins (Hamburg) (Fig. 2) and Jes Rust and Florian with comments on larval evolution within the Myr­ Schroder (Gottingen University) (Fig. 3-6). meleontoidea. Psyche 85: 265-274. I am particularly grateful to my wife Inger Andersen Heilmann-Clausen, C. (1995) Palreogene aflejringer for her interest and financial support over many years. over Danskekalken. Pages 69-114 in Nielsen, O. B.: The major part of this work was carried out when I Danmarks geologi fra Kridt til i dag. (In Danish). stayed as a guest at the Zoological Museum, University Aarhus Geokompendier 1: 1-290. of Copenhagen. Hong, Y.-C. (1982) [Neuroptera]. Pp. 155-156 in: [Mesozoic fossil insects of Jiuquan Basin in Gansu Province] (in Chinese). 187 pp. Geological Publish­ References ing House, Beijing. Hong, Y.-C. (1983) [Middle Jurassic fossil insects in Andersen. N. M. (1998) Water striders from the north China] (in Chinese). 223 pp. Geological Pub­ Paleogene of Denmark with a review of the fossil lishing House, Beijing. record and evolution of semiaquatic bugs (Hemiptera, Kriiger, L. (1923) Neuroptera succinica baltica. Die im Gerromorpha). Biologiske Skrifter 50: 1-157. baltischen Bernstein eingeschlossenen Neuroptera Andersen, N. M. & Andersen, S. (1996) [Giant ants and des Westpreussischen Provinzial-Museums (heute other 'danekrre' from the ash series of the Limfjord. Museum fiir Naturkunde und Vorgeschichte) in The study of an extinct insect fauna from the Danzig. Stettiner Entomologische Zeitung 84: 68-92. Paleogene of Denmark] (in Danish). Naturens Larsson, S. G. (1975) Palaeobiology and mode of burial Verden, 1996 (11/12); 417-432. of the insects of the Lower Eocene Mo-clay of Den­ Ansorge, J. (1996) Insekten aus dem oberen Lias von mark. Bulletin of the Geological Society of Denmark Grimmen (Vorpommern, Norddeutscbland). Neue 24: 193-209. Palaontologische Abhandlungen, 1996 (2): 132 pp. + Larsson, S. G. (1978) Baltic amber - a palaeobiological 17 pI. study. Entomonograph 1: 1-192. Archibald, B. (1998a). The Eocene insectfauna ofQuil- MacLeod, E. G. (1964) Comparative morphological 438 Andersen, S. INSECT SYST. EVOL. 32:4 (2001)

studies on the head capsule and cervix of larval Denmark. Bulletin of the Geological Society of Neuroptera (Insecta). Ph.D. Thesis, Harvard Univer­ Denmark 32: 43-65. sity, Cambridge. Riek, E. F. (1956) A re-examination of the mecopteroid MacLeod, E. G. (1970) The Neuroptera of the Baltic and orthopteroid fossils (Insecta) from the Triassic amber. 1. Ascalaphidae, Nymphidae, and Psychopsi­ beds at Denmark Hill, Queensland, with descriptions dae. Psyche 77: 147-180. of further specimens. Australian Journal of Zoology Makarkin, V. N. (1991) Miocene Neuroptera from 4: 98-110, text-fig. 1-5, pI. 1-2. Northern Caucasus and Sikhote-Alin. (In Russian.) Rust, J. (1998) Biostratinomie von Insekten aus der Fur­ Paleontologicheskii Zhurnal, 1991: 57-68. Formation von Danemark (Moler, oberes PaleozanJ Makarkin, V. N. (1997) Fossil Neuroptera of the Lower unteres Eoziin). Paliiontologisches Zeitschrift 72: 41- Cretaceous of Baisa, East Siberia. Part 4: Psy­ 58. chopsidae. Beitriige zur Entomologie 47: 489-492. Rust. J. (1999) Systematik, Paliiookologie, Biogeogra­ Mansell, M. W. (1992) The systematic position of phie und Biostrationomie fossiter lnsekten aus dem Nemopteridae (Insecta: Neuroptera: Myrmeleotoi­ PaleoziinlEoziin von Diinemark und die Evolution der dea). pp. 233-241 in Canard, M. et al.: Current re­ Insekten im iiltesten Tertiiir. Unpublished Habili­ search in neuropterology. Proceedings of the Fourth tationsschrift, Georg-August-Universitiit, Gottingen. International Symposium on Neuropterology, SchrOder, F. (1999) Ursprungliche Dipteren aus dem Bagneres-de-Luchon, France, 1991. Sacco, Toulouse. Aittertiiir der Fur- und @lst-Fomzation von Diine­ Martynova, O. M. (1947) Kalligrammidae from the mark (Moler, oberes Paleoziinlunteres EOziin). Un­ Jurassic shales of Karatau (in Russian). Doklady published Diplomarbeit, Georg-August-Universitat, Akademii Nauk SSSR 58: 2055-2058. Gottingen. Martynova, O. M. (1949) Mesozoic lacewings (Neu­ Schltiter, T. (1986) The fossil Planipennia: a review. Pp. roptera) and their bearing on concepts of phylogeny 103-111 in Gepp, J. et al.: Recent research in neu­ and systematics of the order. (In Russian). Trudy ropterology. Proceedings of the Second International Palaeontologicheskogo Instituta 20: 150-170. Symposium ofNeuropterology, Hamburg, 1984. Graz. Martynova, O. M. 1954. Neuroptera from Cretaceous Tillyard, R. J. (1918) Studies in Australian Neuroptera. beds of Siberia. (In Russian). Doklady Akademii No.6. The family Psychopsidae, with descriptions of Nauk SSSR 94: 1167-1169. new genera and species. Proceedings of the Linnean Martynova, O. M. (1962) Order Neuroptera. Pp. 272- Society of New South Wales 43: 750-786. 282. In: Rohdendorf, B. B.: Textbook of Palaeontol­ Tillyard, R. J. (l919a) Mesozoic insects of Queensland. ogy, 9: Arthropoda - Tracheata and Chelicerata. (In No.5. Mecoptera, the new order Paratrichoptera, and Russian). Akademiia nauk USSR. [English transla­ additions to the Planipennia. Proceedings of the tion 1991 in Rohdendorf, B. B.: Fundamentals ofPaL­ Linnean Society of New South Wales 44: 194-212. eontology, Vol. 9: Arthropoda - Tracheata, Cheli­ Tillyard, R. J. (1919b) The panorpoid complex. Part 3: cerata. 894 pp. Smithsonian Institution Libraries and the wing venation. Proceedings of the Linnean The National Science Foundation, Washington, D.C.] Society of New South Wales 44: 533-718. New, T. R. (l989a) The Psychopsidae (Insecta: Neurop­ Tillyard, R. J. (1922) Mesozoic insects of Queensland. tera) of Australia and the Oriental Region. Inver­ No.9. Orthoptera, and additions to the Protorthop­ tebrate Taxonomy 2 (1988): 841-883. tera, Odonata, Hemiptera and Planipennia. Procee­ New, T. R. (1989b) Planipennia, Lacewings. Handbuch dings of the Linnean Society of New South Wales 47: der Zoologie, Vol. 4 (Arthropoda: Insecta), Pars 30: 447-470. l32 pp. Tillyard, R. J. (1923) Mesozoic insects of Queensland. Oswald, J. D. (1993) Phylogeny, taxonomy, and bio­ No. 10. Summary of the Upper Triassic insect fauna geography of extant silky lacewings (Insecta: Neu­ of Ipswich, Q. (With an appendix describing new roptera: Psychopsidae). Memoirs of the American Hemiptera and Planipennia). Proceedings of the Entomological Society 40: 1-65. Linnean Society of New South Wales 48: 481-498. Oswald, J. D. (1995) Revision of the southeast Asian Tjeder, B. (1960) Neuroptera - Planipennia. The lace­ silky lacewing genus Balmes (Neuroptera: Psychop­ wings of southern Africa. 3. Family Psychopsidae. sidae). Tijdschrift voor Entomologie l38: 89-101. Pp. 164-209 in Hanstrom, B. et aI.: South African Panfilov, D. V. (1968) Kalligrammatids (Neuroptera: Animal Life, 7. Stockholm. Kalligrammatidae) from Jurassic deposits of the Withycombe, C. L. (1925) Some aspects of the biology Karatau. Pp. 166-174, fig. 1-8. In: Rohdendorf, B. B.: and morphology of the Neuroptera. With special ref­ Jurassic insects of the Karatau. (In Russian). Nauka, erence to the immature stages and their possible phy­ Moscow. logenetic significance. - Transactions of the Panfilov, D. V. (1980) New representatives of lacewings Entomological Society of London, 1924: 303-411. (Neuroptera) from the Jurassic of Karatau. Pp. 82- Zalessky, G. M. (1953) New localities of Cretaceous Ill, fig. xx in Dolin, V. et al.: Mesozoicfossil insects. insects in the Volga, Kazakhstan, and Transbaikal (In Russian). Naukova Dumka, Kiev. regions. (In Russian). Doklady Academii Nauk SSSR Pedersen, G. K. & Surlyk, F. (1983) The Fur Formation: 89: 163-166. a late Paleocene ash-bearing diatomite from northern

Revised manuscript accepted August 2001.

Bibliography of the Neuropterida

Bibliography of the Neuropterida Reference number (r#): 10152

Reference Citation: Andersen, S. 2001 [2001.??.??]. Silky lacewings (Neuroptera: Psychopsidae) from the Eocene-Paleocene transition of Denmark with a review of the fossil record and comments on phylogeny and zoogeography. Insect Systematics and Evolution 32:419-438.

Copyrights: Any/all applicable copyrights reside with, and are reserved by, the publisher(s), the author(s) and/or other entities as allowed by law. No copyrights belong to the Bibliography of the Neuropterida.

Notes:

File: File produced for the Bibliography of the Neuropterida (BotN) component of the Lacewing Digital Library (LDL) Project, 2015.