Introduction

Zorapterans or “Angel Insects” belong to the most enigmatic hexapod groups with respect to their phylogenetic origin, and a term like “Zoraptera-problem” (see Kristensen 1981: “Strepsiptera-problem”) would not be far-fetched. Not less than 11 different systematic placements (see review in Engel and Grimaldi 2002) were discussed since the group was introduced by Silvestri (1913). Silvestri knew only apterous forms by that time, but winged specimens had been already found in 1920. It became clear that zorapterans are in fact characterized by wing dimorphism (Caudell 1920).

The group presently comprises 32 described extant species and 6 species known as fossils (Engel and Grimaldi 2002). The first extinct representative of the group to be discovered was an apterous female of Zorotypus palaeus Poinar 1988 from Dominican amber (Middle Miocene; Poinar 1988; Iturralde-Vinent and MacPhee 1996). Since then, two more species from the Miocene and four from middle Cretaceous deposits were described (Engel and Grimaldi 2000, 2002). Traditionally and in extant classifications (see Engel and Grimaldi 2002), all extant zorapteran species (and most fossils) are placed in one family and one genus (Zorotypidae, Zorotypus). However, a division into three genera was proposed by Kukalova-Peck and Peck (1993) based on differences in the wing venation.

Zorapterans are largely restricted to the tropical and subtropical regions, but Zorotypus hubbardi Caudell 1918 has expanded its range in the United States as far north as Indiana, Iowa and Illinois (Riegel 1963) and two species occur in Tibet (M. Engel, personal communication). The group is probably absent from Australia, but represented in New Zealand and on the Easter Islands (distribution see Weidner 1969; Choe 1989).

The adults are 2–3 mm long and are usually found in colonies under bark of rotting logs together with nymphs. A certain stage of decay with a relatively intact cambium seems to be most suitable and Wilson (1959) defined a ‘zorapteran stage’ in the decomposition of logs in New Guinean rain forests.

The monophyly of Zoraptera is generally agreed upon (see Engel 2003a), but their placement within the neopteran groups is far from being resolved. A sister-group relationship with Embioptera was proposed by Minet and Bourgoin (1986), Engel and Grimaldi (2000) and Grimaldi (2001). A basal placement within Paraneoptera, i.e. a sistergroup relationship with Acercaria was suggested by Hennig (1953, 1969), and this was also considered as a working hypothesis by Kristensen (1975, 1981, 1991) (see also Willmann 2003). A placement close to Dictyoptera was proposed by Wheeler et al. (2001) based on an analysis of morphological data, and 18S- and 28S-rDNA sequences. A further alternative suggested by Kukalová-Peck and Peck (1993) is the placement within a lineage called “Blattoneoptera” (= Dictyoptera + Dermaptera + Grylloblattodea). Further recently discussed options are a sister-group relationship with Eumetabola (e.g. Beutel and Gorb 2001) or with Endopterygota (Rasnitsyn 1998).

One of the main problems underlying the unclear systematic position of Zoraptera is likely the lack of morphological and anatomical data. Kristensen (1981) pointed out that this is probably the “least known insect order” in terms of morphology. Since then, little progress has been made except for a short treatment of thoracic structures by Rasnitsyn (1998). Consequently, the major goal of the present contribution is to improve the knowledge of structural features. Head structures were chosen as they generally offer a broad spectrum of characters (sensory organs, mouthparts, muscles, endoskeleton, fore gut, brain). Structural features found in wingless forms are compared to the conditions in alate specimens, and a reconstruction of the groundplan of the zorapteran head is attempted. A comprehensive table of the head musculature of neopteran insects is presented. Finally, the results are discussed with respect to their phylogenetic implications. It is well understood that a reliable placement of an enigmatic group like Zoraptera has to be based on a cladistic evaluation of a broad spectrum of taxa and characters. However, such an analysis would have been clearly beyond the scope of this primarily morphological contribution.

Materials and methods

The morphological investigations were largely based on wingless forms of Zorotypus hubbardi Caudell 1918, as only two alate specimens (preserved in 70% ethanol) could be used for microtome sectioning or dissection (loan of Smithsonian Institution, Washington D.C.).

Wingless specimens were fixed in FAE (formaldehyde–ethanol–acetic acid—3:6:1) and stored in ethanol (70). Slide preparations of mouthparts and antennae were made with Euparal as embedding medium. Skeletal preparations of the head were made after maceration in KOH. For the reconstruction of musculature, digestive tract, endoskeleton and brain serial cross sections, frontal sections and longitudinal sections were made. The specimens were embedded in HistoResin and cut at 3 μm with a Microm microtome (HM 360). The sections were stained with methylene blue and acid fuchsine.

Other specimens were cleaned with ultrasonic sound and sputter-coated with gold for scanning electronic microscopy. Pictures were taken with an FEI scanning electronic microscope (XL 30 ESEM) (FSU Jena) and a Cambridge Stereoscan MK2 (Universität Tübingen).

Drawings were made with a camera ludica or with an ocular grid and processed and evaluated with Adobe Photoshop and Macromedia Freehand software. AnalySIS software was used for the documentation of serial sections and slide preparations. The 3-dimensinal reconstruction was carried out with MAYA software.

The muscular terminology follows v. Kéler (1963). For comparative purposes, specimens of Ectobius sylvestris (Poda, 1761), Periplaneta americana (Linné, 1758) (Blattodea), Sphodromantis sp. (Mantodea), Epipsocus lucifugus (Rambur, 1842) and Trogium pulsatorium (Linné, 1758) (Psocoptera) were examined.

Head morphology of apterous specimens

External head capsule

The head is not retracted into the prothorax. It articulates with a well-developed pair of ventrolateral cervical sclerites on its posterolateral margin. The foramen occipitale is distinctly narrowed by the gula (see below). The head is orthognathous and about as long as broad posteriorly, and triangular in dorsal view (Fig. 1a, b). The thin cuticle is largely unpigmented. Longer and shorter setae are distributed as shown in Figs. 1a and b. Compound eyes and ocelli are absent. The coronal, frontal, genal and postoccipital sutures are absent (Fig. 1a). The clypeus is distinctly divided into a largely unsclerotized anterior anteclypeus and a posterior postclypeal area, which is strongly narrowed between the antennal articulation areas and posteriorly fused with the frons (Figs. 2a, 5a). Both clypeal parts are separated by a shallow rim externally and by a distinct, sclerotized transverse ridge internally (Fig. 5a: tcr). The maxillary grooves between the labium and the genal region are well developed. The broad gular area (Figs. 1b, 2b) is less strongly sclerotized than other parts of the head capsule. It is laterally limited by a fold, parallel to a distinct hypostomal ridge.

Fig. 1
figure 1

Zorotypus hubbardi, head, wingless form. a Dorsal view. b Ventral view. aart antennal articulatory area, ata anterior tentorial arm, c cardo, cgld cervical gland duct, dta dorsal tentorial arm, foc foramen occipitale, ga galea, gl glossa, gu gula, hyr hypostomal ridge, lbr labrum, lc lacinia, md mandible, pcl postclypeal area, pg palpiger, pgl paraglossa, pl palpus labialis, pmt prementum, pmx palpus maxillaris, psm postmentum, pta posterior tentorial arm, ptg posterior tentorial grooves, st stipes, tb tentorial bridge

Full size image
Fig. 2
figure 2

Zorotypus hubbardi, head, wingless form, SEM micrographs. a Dorsolateral view. b Ventrolateral view. acl anteclypeus, arm articulatory area of maxilla, c cardo, gl glossa, gu gula, hy hypopharynx, lbr labrum, lc lacinia, md mandible, pl palpus labialis, pm postmentum, prm prementum, pmx palpus maxillaris, st stipes

Full size image

Musculature: a transverse muscle with unclear homology (Mxy) is attached to the vertex (or posterior frontal area) on both sides. The muscle separates the anterior margin of the brain from the posterior components of M. frontopharyngalis posterior (M. 46) (Figs. 5a, 6a: mxy).

Cephalic endoskeleton

The postoccipital ridge is narrow. A low internal ridge corresponds with the hypostomal suture internally. The frontoclypeal suture corresponds with a distinct internal transverse ridge. The tentorium is well developed. All parts are sclerotized and distinctly flattened. The fissure-shaped posterior tentorial grooves are posteriorly continuous with the hypostomal sutures (Fig. 1b: hyr). The anterior grooves are visible as distinct darkened fissures close to the dorsal (secondary) mandibular articulation. The short posterior arms are connected by the broad and straight tentorial bridge (Fig. 5a). The anterior arms are the largest part of the tentorium. They are flattened and distinctly broadened towards their anterior origin. The flat and straight dorsal arms are dorsally attached to the head capsule by means of numerous muscle fibres. An accessory anterior bridge formed by mesally directed arms (laminatentorium; “perforated tentorial bridge”) is absent.

Labrum

The labrum is roughly triangular and about as long as broad (Figs. 1a, 2a, 3a); it is rounded posterolaterally and anteromedially; the base is distinctly narrowed and connected with the anterior margin of the clypeus by a membranous fold. Setae are inserted along the anterior margin and anterodorsally. Fine transverse lines are present on the dorsal surface. The tormae, which serve as attachment area of M. frontoepipharyngalis (M. 9) are short.

Fig. 3
figure 3

Zorotypus hubbardi, head, wingless form. a Labrum, ventral view. b Antenna. c Mandible, ventral view. d. Maxilla, dorsal view. con condyle of primary mandibular joint, eph epipharynx, ga galea, lc lacinia, mo mola, ped pedicellus, pm1 palpomere 1, pst prostheca, sc scapus, 7 attachment area of M. labroepipharyngalis, 8 M. frontolabralis, 9 M. frontoepipharyngalis, 17 M. tentoriocardinalis, 18 M. tentoriostipitalis, 19 M. craniolacinialis, 20 M. stipitolacinialis, 22/23 M. stipitopalpalis ext./int., 24 M. palpopalpalis max. prim

Full size image

Musculature (Figs. 3a, 5a, 7a, b): M. labroepipharyngalis (M. 7), strongly developed—origin (O): central region of external wall of labrum; insertion (I): paramedially on anterior epipharynx; M. frontolabralis (M. 8)—O: medially on posterior frontal region; I: external wall of labrum, between labral base and origin of M. labroepipharyngalis (M. 7); M. frontoepipharyngalis (M. 9)—O: posterior frontal region, between M. frontolabralis (M. 8) and M. frontohypopharyngalis (M. 41); I: posterolateral edge of labrum, on short tormae; M. epistomalabralis (M. 10)—absent.

Antenna

The antenna is composed of nine antennomeres (adults) (Fig. 3b). The large membranous articulatory areas lie closely together on the dorsal side of the head, enclosing the narrow clypeal area between them (Fig. 1a). The scapus is elongate, approximately three times longer than wide, with a distinct constriction close to its base and a small anterior condyle. The pedicellus and flagellomere 1 are as broad as the scapus base, and together slightly shorter than the scapus. The Johnston’s organ is poorly developed, with only few scolopidia. Flagellomeres 2–7 are distinctly longer and broader than antennomeres 2 and 3 and approximately egg-shaped; the stalk-like, narrowed basal part is inserted in the articulatory membrane of the preceding segment. All flagellomeres are covered with setae (sensilla) of different length. The distribution is shown in Fig. 3b.

Musculature (Fig. 6a, b): very strongly developed, M. tentorioscapalis anterior (M. 1)—O: anterior part of anterior tentorial arm and lateral side of dorsal arm; I: anterolaterally on scapal base; M. tentorioscapalis posterior (M. 2)—O: mesally on dorsal tentorial arm; I: posteriorly on scapal base; M. tentorioscapalis lateralis (M. 3)–absent; M. tentorioscapalis medialis (M. 4)—O: anterior tentorial arm, posterior to origin of dorsal arm; I: anteromesally on scapal base; M. scapopedicellaris lateralis (M. 5)—O: posterior wall of scapus; I: posteriorly on the base of the pedicellus; M. scapopedicellaris medialis (M. 6), distinctly larger than M. 5—O: large parts of the basal part of the scapus; I: anteriorly on the base of the pedicellus.

Mandibles

The mandibles are strongly sclerotized, almost triangular with a rounded lateral margin, and distinctly asymmetric (Figs. 3c, 4a). They are largely covered by the labrum in their resting position. The ventral side is adjacent to the maxillae. Five teeth are present on the distal part of both mandibles; the two apical teeth of the left mandible are equally sized and lie above each other; they are separated from the smaller subapical tooth by a notch; the moderately sized subapical tooth is separated from the large following intermediate tooth by a deep incision; the triangular proximal tooth is posteriorly adjacent with the molar area, which is equipped with tubercles in its anterior portion. The posterior part of the left mola lacks tubercles or ridges. The dorsal apical tooth of the right mandible is distinctly shorter than its ventral counterpart; the subapical tooth is large, triangular and pointed; it is followed by a very small posteriorly directed tooth and by a triangular proximal tooth; the tuberculate grinding surface of the anterior molar part is more distinct than on the left mandible. A brush-like prostheca is moveably articulated with the ventral surface of the left mandible, lateral to the large intermediate tooth (Fig. 3c). It is absent on the right mandible.

Fig. 4
figure 4

Mouthparts, SEM micrographs. a Right mandible, ventral view. b Apex of maxilla, mesal view. c Labium and hypopharynx, lateral view. ga galea, gl glossa, hy hypopharynx, lc lacinia, pg palpiger, pgl paraglossa, pmt prementum

Full size image

Musculature (Figs. 5a, 6c): M. craniomandibularis internus (M. 11), by far the largest muscle of the head capsule, with an almost vertical anterior subcomponent—O: dorsally (largest part), laterally and ventrolaterally from the posterior head capsule; I: strongly developed adductor tendon; M. craniomandibularis externus (M. 12)—O: posterolaterally from the head capsule; I: abductor tendon; M. hypopharyngo-mandibularis (M. 13)—O: ventral side of anterior tentorial arm; I: ventrobasal surface of the mandible; M. zygomaticus mandibulae (M. 14)—absent.

Fig. 5
figure 5

Zorotypus hubbardi, head. a Wingless form, sagittal section, cell bodies of the part of the brain covering M. cranioandibularis int. omitted. b Alate form, head in lateral view. ata anterior tentorial arm, acl anteclypeus, c cardo, cbr cell body rind, cer cerebrum, ce compound eye, cgl cervical gland, cgld cervical gland duct, ga galeae, gl glossa, lbr labrum, lc lacinia, lm longitudinal muscle, mo mola, mxy transverse muscle of head capsule, oc ocellus, oes oesophagus, pl palpus labialis, pmx palpus maxillaris, pta posterior tentorial arm, rm ring muscle, sc scapus, sgl salivary glands, soec suboesophageal complex, st stipes, tb tentorial bridge, tcr transverse clypeal ridge, 7 M. labroepipharyngalis, 8 M. frontolabralis, 9 M. frontoepipharyngalis, 11 M. craniomandibularis int., 17/18 Mm. tentoriocardinalis/-stipitalis, 28 M. submentopraementalis, 29/30 Mm. tentoriopraementalis inf./sup., 34 M. praementopalpalis ext., 41 M. frontohypopharyngalis, 42 M. tentoriohypopharyngalis, 43a, b M. clypeopalatalis, 45 M. frontobuccalis ant., 46 M. frontobuccalis post., 52 M. tentoriopharyngalis

Full size image
Fig. 6
figure 6

Serial sections, ca. 90° to longitudinal body axis. a dorsal region, frontal ganglion. b Mandibular region, with cibarium. c Maxillary region. cbr cell body rind, cer cerebrum, gf ganglion frontale, md mandible, mx maxilla, mxy transverse muscle of head capsule, ph pharynx, rm ring muscle, soec suboesophageal complex, tb tentorial bridge, tm transverse muscle, 1, 2, 4 Mm. tentorioscapalis ant., post., med., 7 M. labroepipharyngalis, 8 M. frontolabralis, 9 M. frontoepipharyngalis, 11 M. craniomandibularis internus, 12 M. craniomandibularis externus, 17/18 Mm. tentoriocardinalis/-stipitalis, 19 M. craniolacinialis, 20 M. stipitolacinialis, 29 M. tentoriopraementalis inf., 43a, b M. clypeopalatalis, 46 M. frontobuccalis post., 47 M. frontobuccalis lat., 52 M. tentoriopharyngalis

Full size image

Maxillae

The maxillae are well developed, very slightly asymmetric (lacinia), and inserted in a deep maxillary groove. The articulatory membrane is well developed (Figs. 1b, 2b). The distribution of setae is shown on Figs. 1b and 3c. The cardo is approximately triangular and divided into a mesal proxicardo and a lateral disticardo, which is broadly connected with the stipes; the margin of both parts is reinforced by a distinct ridge; a condyle formed at the border of both subunits articulates with the head capsule; an elongated apodeme for attachment of M. craniocardinalis externus (see below) inserts on the basal edge of the disticardo (Fig. 3d). The stipes is not subdivided into a basistipes and mediostipes and it is also fused with the slender lacinia. The apical part of the left lacinia is divided into two teeth (Fig. 4b). The subapical tooth of the right lacinia is smaller and more widely separated from the apical tooth. The mesal edge of the laciniae is smooth and completely devoid of setae or spines (Fig. 4b). The apically rounded semimembranous galea is broadly attached to the anterior margin of the mediostipital area (Fig. 3d). The apical part is set with a group of mesally directed strong spines and a comb of long, apically curved setae (Fig. 4b). The palpifer is not distinctly separated from the stipes. The palp is five-segmented; the short proximal palpomere is about as wide as long; palpomeres two and three are narrowed at their base and almost four times as long as palpomere one; palpomere four is slightly longer than one, with a strongly narrowed, stalk-like proximal part and an almost globular distal part; the apical palpomere is basally distinctly narrowed and distinctly larger than palpomere four; it is about as long as palpomeres one to three together; it is densely covered with medium-length setae and densely arranged short setae are present at the apex.

Musculature (Figs. 3c, 6c): M. craniocardinalis externus (M. 15), almost vertical, slender muscle—O: dorsolaterally on posterior head capsule, between M. craniomandibularis externus (M. 12) and M. craniolacinialis (M. 19; see above); I: laterally on the base of the cardo; M. craniocardinalis internus (M. 16)—absent; M. tentoriocardinalis (M. 17)—O: anterior tentorial arm, close to origin of M. tentoriostipitalis; I: laterally on the inner surface of the cardo; M. tentoriostipitalis (M. 18)—O: bipartite, ventral side of anterior tentorial arm; I: ventrally on mesal edge of stipes, proximad to attachment of M. craniolacinialis; M. craniolacinialis (M. 19)—O: head capsule between M. craniocardinalis (M. 15) and the dorsal component of M. craniomandibularis internus (M. 11); I: basal edge of lacinia by means of tendon; M. stipitolacinialis (M. 20), well developed—O: laterally on stipital base; I: base of lacinia, close to insertion of M. 19; M. stipitogalealis (M. 21), two bundles with single insertion—O: posterior edge of basistipes and mesal edge of mediostipes; I: base of galea by means of a short thin tendon; M. stipitopalpalis externus (M. 22)—O: ventrally from stipes, close to mesal edge; I: posteriorly on the base of palpomere 1; M. stipitopalpalis internus (M. 23)—O: anterior to M. stipitopalpalis externus; I: anteriorly on the base of palpomere 1; M. palpopalpalis maxillae primus (M. 24)—O: anteriorly from the basal part of palpomere 1; I: base of palpomere 2; M. palpopalpalis secundus (M. 25)—O: basal part of palpomere 2; I: base of palpomere 3; M. palpopalpalis tertius (M. 26)—O: basal part of palpomere 3; I: base of palpomere 4; Musculus palpopalpalis quatrus (M. 27)—O: posteriorly on the basal part of palpomere 4; I: anteriorly on the base of palpomere 5.

Labium

The comparatively small postmentum is separated from the foramen occipitale by the large gular area (Figs. 1b, 2b). It is inserted between the maxillary articulatory areas and divided into a very short submentum and a quadrangular mentum. The setae are distributed as shown in Figs. 1b and 4c. The prementum is about as large as the postmentum and medially divided by a rim, which corresponds with a median internal ridge (Fig. 2b). The palpiger is well developed. The palp is three-segemented. The proximal palpomere is elongate, about 4.5 times as long as wide; a small proximal piece is distinctly separated from the rest, thus suggesting the presence of an additional palpomere. Palpomere 2 is small, about as long as wide; the base is distinctly narrowed and stalk-like. The apical palpomere is large and broadenend, with a strongly narrowed basal part. The apex is nearly truncate, with a slightly sinuate anterior edge and rather densely set with short setae. The distribution of the setae is shown in Figs. 1b, 2b and 4c. Well-developed, basally fused glossae and paraglossae are attached to the prementum above the palpigers (Figs. 2a, 4c). The distal part of the glossae is membranous, smooth and wrapped around the anterior hypopharyngeal margin. The laterodistal part of the paraglossae is closely adjacent with the lateral hypopharyngeal margin. Both parts are densely set with setae, especially in the anterior region (except for glossal apices).

Musculature (Figs. 5a, 6c, 8): M. submentopraementalis (M. 28), composed of two subcomponents, M. 28a—O: medially on posterior gula; I: posterior margin of median premental ridge; M. 28b—O: medially on hind margin of postmentum; I: together with M. 28a; M. tentoriopraementalis inferior (M. 29)—O: posterior tentorial arm and base of tentorial bridge; I: posterior margin of prementum, lateral to M. submentopraementalis a; M. tentoriopraementalis superior (M. 30)—O: tentorial bridge, close to the origin of M. 29; I: salivary sclerite; M. praementoparaglossalis (M. 31)—O: median apodeme of prementum; I: external wall of paraglossa; M. praementoglossalis (M. 32)—O: ventral wall of prementum; I: ventrally on hind margin of glossa; M. praementopalpalis internus (M. 33)—O: median ridge of prementum; I: mesally on the base of palpomere 1; M. praementopalpalis externus (M. 34), bipartite muscle, with a large main subcomponent and slender lateral subcomponent—O: on the lateral wall of the prementum and a large premental apodeme, which reaches the anterior mental region posteriorly (main subcomponent); I: laterally on the base of palpomere 1; M. palpopalpalis labii primus (M. 35)—O: mesal wall of palpomere 1; I: laterally on the base of palpomere 2; M. palpopalpalis labii secundus (M. 36)—O: palpomere 2; I: basally on palpomere 3.

Epipharynx

The epipharynx (Fig. 3a) is semimembranous and longitudinally divided into an anterior part (preepipharynx), which forms the roof of the open preoral cavity (=corrosium [working space for mandibular incisive] and molarium [working space of molae] sensu Kéler 1966), and a posterior part, which forms a short closed tube (prepharynx) together with the mesal mandibular base (anterior part) and the posthypopharynx (see below; =cibarium sensu v. Kéler 1966) (Fig. 5a). The anterior part of the preepipharynx is flat (=ventral wall of labrum) with indistinct folds lateral to the attachment area of M. labroepipharyngalis and a short and indistinct median bulge with very short microtrichia. Rows of extremely short spines or microtrichia enclose a distinct median rim of the posterior part of the preepipharynx. The prepharynx is approximately U-shaped in cross section (posterior part) (Fig. 7b); its anterior opening is the functional mouth.

Fig. 7
figure 7

Serial sections, ca. 90° to longitudinal axis of head. a Anterior buccal cavity. b Posterior cibarium. hy hypopharynx, md mandible, mdb mandibular base, mri median ridge of prementum, mx maxilla, mxz retractor of salivary sclerite, pmt prementum, sal salivarium, salscl salivary sclerite, tm transverse muscle, 1, 4 Mm. tentorioscapalis ant., med.; 5 M. scapopedicellaris lateralis, 8 M. frontolabralis, 9 M. frontoepipharyngalis, 18 M. tentoriostipitalis, 21 M. stipitogalealis, 22/23 M. stipitopalpalis ext./int., 24 M. palpopalpalis max. prim., 28 M. submentopraementalis, 29, 30 M. tentoriopraementalis inf., sup., 32 M. praementoglossalis, 33, 34 Mm. praementopalpalis int., ext., 37 M. hypopharyngosalivaris, 38 M. praementosalivaris ant., 40 M. annularis salivaris, 43a, b M. clypeopalatalis, 47 M. frontobuccalis lat

Full size image

Musculature (Figs. 3a, 5a): M. clypeopalatalis (M. 43), very strongly developed, composed of two subcomponents; M. 43a—O: anterolateral part of postclypeal area; I: medially on preepipharyngeal roof; M. 43b, a series of 7–8 bundles—O: posterior to M. clypeopalatalis a (M. 43a); I: successively on the preepipharyngeal and prepharyngeal roof. The bundles of M. clypeopalatalis (M. 43) are separated by an exceptionally strong transverse musculature (Fig. 7b: tm).

Salivarium and salivary glands

The salivarium is present as a flat and very narrow pocket between the hypopharynx and the anterior labium (Fig. 5a). Its ventral wall and the anterolateral edges are sclerotized (salivary sclerite, see above). It is connected with the well-developed salivary gland posteriorly.

Musculature (Fig. 7a): M. hypopharyngosalivarialis (M. 37), a thin muscle, O: – posterolateral edge of the hypopharynx; I: dorsally on the salivarium; M. praementosalivarialis anterior (M. 38)—O: anterolateral wall of the prementum, anterior to the smaller subcomponent of M. praementopalpalis lateralis (M. 34); I: ventromedially on the salivary sclerite; M. praementosalivarialis posterior (M. 39)—O: posterolateral wall of the prementum; I: laterally on the salivary sclerite; M. annularis salivarii (M. 40)—probably represented by a small transverse muscle connecting the anterolateral edges of the salivary sclerite.

M. tentoriopraementalis superior (M. 30) is attached to the salivary sclerite (see above) and also an additional slender muscle (Fig. 7b: mxz), which originates together with M. 28a.

Hypopharynx

The hypopharnyx forms a functional and structural unit with the anterior labium, but is internally separated from it by the narrow salivarium (see below; Figs. 4c, 5a, 7a, b). Longitudinally it can be divided into a tongue-like anterior part (prehypopharynx), which lies above the prementum and between the mediodistal part of the maxillae, and a posterior part (posthypopharynx), which is part of the prepharyngeal tube (see above). The dorsal surface of the prehypopharynx is semimembranous; it is flat anteriorly and covered with rows of very fine microtrichia; a median longitudinal bulge, also covered with fine microtrichia, is present on the posthypopharynx. The lateral edge of the posterior prepharynx is reinforced by distinctly sclerotized flat suspensoria.

Musculature (Figs. 5a, 7b): M. frontohypopharyngalis (M. 41)—O: posterior frontal area, lateral to the origin of M. frontoepipharyngalis (M. 9); I: apex of the suspensorium at the anatomical mouth; M. tentoriohypopharyngalis (M. 42)—O: tentorial bridge; I: ventromedially on the posterior hypopharynx.

Pharynx and oesophagus

The lumen of the pharynx is rather narrow (Fig. 6c). Dorsal, ventrolateral and ventral folds serve for muscle attachment. A dorsal deep fold is present immediately posterior to the anatomical mouth (Fig. 5a). The pharynx is posteriorly continuous with the wide and strongly folded oesophagus (Figs. 5a, 8b, d).

Fig. 8
figure 8

Zorotypus hubbardi, brain, suboesophageal complex and fore gut, wingless form. a Dorsal view; b Lateral view; c Anterolateral view; d Posterolateral view. an antennal nerve, con circumoesophageal connective, dcer deutocerebrum, oes oesophagus, ph pharynx, pcer protocerebrum, soec suboesophageal complex

Full size image

Musculature (Figs. 5a, 6c): M. clypeobuccalis (M. 44)—probably absent; it cannot be fully excluded that M. clypeobuccalis is represented by the posteriormost bundle of M. clypeopalatalis b (M. 43b); however, this appears rather unlikely as this is a very compact group of muscles; M. frontobuccalis anterior (M. 45), a well-developed muscle, separated from M. clypeopalatalis (M. 43) by the frontal connectives—O: anterior part of the frontal area; I: dorsally on the anatomical mouth, anterior to the dorsal fold and between the attachment of M. frontohypopharyngalis (M. 41); M. frontobuccalis posterior (M. 46), a series of bundles—O: posterior frontal area, close to anterior margin of the brain; I: the anterior bundle inserts on the dorsal fold, the following bundles successively on the precerebral part of the pharynx; M. frontobuccalis lateralis (M. 47), a transverse muscle—O: laterally from the head capsule, close to the origin of the anterior tentorial arm and secondary mandibular joint; I: laterally on the anatomical mouth; it cannot be fully excluded that this is an atypical lateral subcomponent of M. frontohypopharyngalis (M. 41); M. tentoriobuccalis anterior (M. 48)—absent; M. tentoriobuccalis lateralis (M. 49)—absent; M. tentoriobuccalis posterior (M. 50)—O: laterally on the tentorial bridge, close to M. tentoriopharyngalis (M. 52); I: ventrolateral folds of the anterior pharynx, below the insertion of M. frontobuccalis posterior (M. 46); M. verticopharyngalis (M. 51)—O: from the posterior head capsule between the posterior part of the brain and the dorsal bundles of M. craniomandibularis internus (M. 11); I: dorsal fold of the posterior pharynx; M. tentoriopharyngalis (M. 52)—O: tentorial bridge; I: ventral fold of the pharynx, below the attachment of M. verticopharyngalis (M. 51); M. transversalis buccae (M. 67), moderately developed, connects the upper edges of the anatomical mouth; a very strong ventral transverse muscle is attached to the suspensorium on both sides; M. anularis stomodaei (M. 68)—a well-developed ring musculature is present over the whole length of the pharynx; it is less regularly arranged at the oesophageal folds which are posteriorly adjacent with the pharynx; M. longitudinalis stomodaei (M. 69)—well-developed longitudinal muscles are present on the dorsal and ventral sides of the pharynx. They are covered by the ring musculature (Figs. 5a, 6c).

Cervical glands

Large asymmetric glands (Fig. 5a: cgl) are present in the neck region between the suboesophageal complex and the oesophagus. They are connected with a sclerotized tube which opens dorsomedially on the posterior head capsule.

Brain and suboesophageal complex

The brain and the suboesophageal complex are very large in relation to the head size (Fig. 8). They are distinctly modified in shape in order to fit very closely between muscles and endoskeletal structures. Distinct recesses are present at the position of the anterior parts of M. craniomandibularis internus, and the dorsal tentorial arms. The protocerebrum is strongly inclined posteriorly. The posterior part is increasingly flattened towards the postoccipital area. Two strongly asymmetric protocerebral lobes reach the foramen occipitale posteriorly. The large corpora pedunculata are distinct. The optic neuropils are reduced and ocellar nerves are absent. The deutocerebrum forms the anteriormost part of the cerebrum and the olfactory lobes and the antennal nerves are strongly developed (Fig. 8a–c). The tritocerebral neuropil is distinct. The commissure is not recognizable as a separate structure. The circumoesophageal connectives are short and massive (Fig. 8b). They are very broadly connected with the large suboesophageal complex, which reaches the anatomical mouth region anteriorly and the ventral neck region posteriorly. The suboesophageal ganglion is almost rectangular in cross section with distinct dorsolateral edges. The connectives linking it with the prothoracic ganglion are long and moderately thick.

Head morphology of winged species

The head capsule and appendages are brownish. Well developed compound eyes with numerous small ommatidia are present posterolaterally (Fig. 5b). The optic neuropils are well developed. The ocelli are arranged as a triangle, two anterior to the compound eyes, and one medially on the anterior frons. Kidney-shaped white mark are not present between the single ocellus and the compound eyes (Fig. 5b) (see Weidner 1970b). Other external and internal structures do not differ from the apterous form.

Discussion

Miniaturisation

With a total size ranging from 1.5 mm to 2.5 mm zorapterans belong to the smaller insects. This arises the question whether structural modifications may be related to size reduction. Miniaturisation can result in structural simplification (Hanken and Wake 1993), but this is rarely the case even in very small insects (Beutel and Haas 1998) or only to a minor degree (Grebennikov and Beutel 2002). Externally, the only derived feature (see below), which may be related to small size is the far-reaching reduction of the head sutures (Figs. 1, 2). The mouthparts, the tentorium and the head musculature are apparently not affected. They are as complete and complex as in most other insects. Internally, however, musculature, endoskeleton and the brain are arranged in a very compact manner as in other small or very small insects (Beutel and Haas 1998; Grebennikov and Beutel 2002). The brain and suboesophageal complex are very large in relation to the size of the head, and form a very compact mass around the pharynx. As in other very small insects, the cell body diameter in the dorsal cell body rind is relatively large (ca. 4–6 μm; Table 1, see Beutel and Haas 1998). The shape of the brain is distinctly modified in order to make it fit perfectly with other internal structures. The morphological upper margin of the protocerebrum reaches the foramen occipitale posteriorly and is strongly asymmetric. In contrast to other small insects (especially larvae; see Beutel and Haas 1998; Grebennikov and Beutel 2002), the brain does not reach into the prothorax. It was pointed out in Beutel et al. (1995) that this is only possible with a wide foramen occipitale and a limited movability of the head. This is clearly not the case in Zoraptera.

Table 1 Head muscles of Zorotypus and representatives of other insect groups (part 1, appendages). Muscle numbers on the left refer to v. Kéler 1963 (also used in the descriptive part). Numbers or letters introduced by the other authors are used in the remaining columns (except for Zorotypus). The presence of 2 or more subunits is either indicated by 2 or more numbers used by the relevant author (e.g. 10–12) or by numbers following a slash (e.g. 2 [=muscle number 2]/2 [=composed of 2 subunits]). Question marks refer to uncertain homology or to uncertain presence or absence. Unusually strong muscles are underlined
Full size table

Plesiomorphies and putative autapomorphies of Zoraptera

The head of Zorotypus is mainly characterized by plesiomorphic features. The orthognathous condition, the complete tentorium (without accessory anterior bridge [“unperforated corpotentorium”], see below), the free labrum, the biting mandibles, maxillae with an articulated galea, a lacinia firmly attached to the stipes and a five-segmented palp, a labium divided into postmentum and a prementum equipped with glossae, paraglossae, palpiger and a three-segmented palp, are very likely groundplan features of Pterygota or Insecta (=Ectognatha). The head musculature is probably close to the groundplan condition of Neoptera (+Odonata). A full set of labral, maxillary and labial muscles is present and the muscles of the salivarium and the pharynx are also largely in agreement with what is found in other neopteran groups (see Table 1). Muscles which belong to the groundplan of Insecta, but are generally absent in Pterygota (M. tergostipitalis, M. tentoriolacinialis, M. postoccipitoglossalis, M. tentoriosalivaris; Chaudonneret 1950; Bitsch 1963; Matsuda 1965) or Pterygota excluding Ephemeroptera (=Odonata +Neoptera) (dorsal tentoriomandibular muscles, second cranial abductor of mandible; Strenger 1953; Staniczek 2000), are also missing in Zorotypus hubbardi. The mandibles are almost exclusively moved by the strongly developed adductor (M. craniomandibularis internus, M. 11) and abductor (M. craniomandibularis externus, M. 12). M. tentoriomandibularis (M. 13) is present but very modestly developed as compared to the other two muscles.

Despite of the large number of plesiomorphic features, the monophyly of Zoraptera is well supported. The eye dimorphism, i.e. the presence or complete absence of compound eyes (and optical lobes: Fig. 8) and ocelli, which is closely linked with the wing dimorphism (Engel 2003b), is one presumptive autapomorphy of the group. Vestiges of the optic apparatus such as externally recognisable pigmentation (Caudell 1918), nerves directed towards the eye region or groups of small black bristles (Weidner 1969) were absent in all specimens of Z. hubbardi examined.

Another likely autapomorphy of Zoraptera is the reduced number of antennomeres (Fig. 3b) as compared to other groups of insects (e.g. Plecoptera, Ensifera, Mantodea, Blattodea, Grylloblattodea, Embioptera, Psocoptera). Adult zorapterans possess seven flagellomeres, whereas only four are present in immature stages (Riegel 1987).

The presence of a moveable prostheca on the left mandible is another potential autapomorphy. This structure was already mentioned by Silvestri (1913) for Z. guineensis and Z. javanicus. Whether it is present in all species has to be clarified in future studies, but its occurrence seems likely (M. Engel, personal communication).

In addition to the wing and eye dimorphism, a sexual dimorphism occurs in Zoraptera. A frontal gland is described for males of some species of Zoraptera (Delamare-Deboutteville 1951; Choe 1989). However, it is absent in Z. hubbardi and therefore probably not a groundplan feature of the entire group. The function and homology of the gland are unclear. It was compared to the frontal porus and the associated frontal glands of termites by Weidner (1970a). However, homology is very unlikely in this case considering the well-founded placement of Isoptera as a subordinate group of Blattodea (most closely related to Cryptocercus; see Klass 2003; Lo 2003).

Another derived condition is the almost complete absence of head sutures in Z. hubbardi (Fig. 1a) and other species (Weidner 1970b). However, as a distinct Y-shaped epicranial cleavage line (frontal and coronal sutures) was depicted by Silvestri (1913) in his description of Z. guineensis (males), this is possibly not a groundplan feature of Zoraptera.

Other derived features of Z. hubbardi are the presence of a transverse muscle of the head capsule (Figs. 5a, 6a: mxy), which is not known from other insects, the presence of a retractor of the salivarium with origin from the gula (Fig. 7b: mxz), the modifications of the brain described above (Fig. 8), and the presence of glands with a sclerotised duct in the neck region (Fig. 5a). Whether these characteristics belong to the groundplan of Zoraptera is unclear, as internal features of other species are unknown. The presence of glands with a sclerotized duct in the posteromedian head region is a highly unusual feature. Dorsal glands are also described for psocopterans (Badonnel 1934; Weber 1938). However, they are connected with the salivarium as are the usual salivary glands in other groups of insects.

The systematic position of Zoraptera

One of the hypothesized systematic placements presently discussed (see review in Engel and Grimaldi 2002 and in Grimaldi and Engel 2005) is a sister-group relationship with Dictyoptera, which was suggested in Wheeler et al. (2001). We found no morphological evidence supporting such a placement of Zoraptera. The only possibly derived feature occuring in both groups is the kidney-shaped frontal spot (Frontomacula, Antennalorgan; Weidner 1969), which is described for winged zorapteran specimens and does also occur in blattodeans and termites. However, it was absent in all alate specimens of Z. hubbardi observed during our study, and may therefore not belong to the groundplan of Zoraptera. Besides this, it is also present in “basal” plecopterans (Weidner 1969), and the homology and function is unclear. The accessory anterior tentorial bridge (“perforated corpotentorium”), which is generally considered as an apomorphy of Dictyoptera, is absent in Zoraptera (see also Kristensen 1975: Fig. 7). Similarities of the antennae of Zoraptera and Isoptera (chemoreceptors, tactile setae, Johnston’s organ) were pointed out by Slifer and Sekhon (1978). However, it is very likely that these are due to parallel evolution as a close relationship between both groups is very unlikely (position of Isoptera see above).

A clade comprising Dictyoptera and Zoraptera was not supported in separate analyses of the morphological and molecular data sets (18S-, 28S-rDNA) in Wheeler et al. 2001 (see Figs. 12, 13, 14), but only in the combined analyses under certain parameter settings (see Wheeler et al. 2001). The proposed morphological synapomorphies of Dictyoptera and Zoraptera (Boudreaux 1979; Wheeler et al 2001: characters 96–98, 100) are not convinicing. A disc-shaped pronotum as it is ascribed to Zoraptera and dictyopterans is quite common in insects, which was also pointed out by the authors (e.g. Grylloblattodea, Dermaptera, Embioptera, Heteroptera, Coleoptera). Forward slanting pleural sutures are probably present in all insects capable of flight. The loss of indirect flight muscles is certainly not a groundplan feature of Zoraptera and Dictyoptera, respectively. The dorsal and dorsoventral muscles of the pterothorax are strongly developed in alate specimens of Z. hubbardi (R. G. Beutel personal observations). Conical, backwards-directed coxae are also commonly found in different groups of insects, and are possibly plesiomorphic (e.g. Zygentoma).

A sister-group relationship between Zoraptera and Blattoneoptera (=Dermaptera + Dictyoptera) was discussed by Kukalová-Peck and Haas (2001). This hypothesis was not supported by the results of our study. It is at least partly based on characters which are not groundplan features of Zoraptera, or on problematic characters, observations and interpretations from Boudreaux (1979) and Rasnitsyn (1998) (see below). As it is pointed out above, the reduced state of the flight apparatus (including phragmata) does not apply to the alate morphs.

Another potential sister group of Zoraptera is Embioptera (Minet and Bourgoin 1986; Engel and Grimaldi 2000; Grimaldi 2001). A considerable number of presumably derived characteristics is shared by both groups. However, none of them is unique to these two taxa and none of them is very specific. Subsociality (e.g. Valentine 1986) is not a very precisely defined character state and may have evolved independently, as this is almost certainly the case in other taxa such as e.g. Campodeidae (Diplura), Cryptocercus (Blattodea) (Deitz et al. 2003), Archipsocopsis (Psocoptera) (Weidner 1972), Phlaeothripidae (Thysanoptera) (Lewis 1973), and Pemphiginae (Aphidiidae) (Carver et al. 1991). The reduced number of flagellomeres is an unspecific reduction, which has doubtlessly evolved several times independently. Besides this, the maximum number of antennomeres in Embioptera is 34. Simplified, paddle wings with a basal fraction zone have also evolved independently in different groups of insects (e.g. Isoptera, Formicidae). It was already pointed out by Hennig (1969) that the reduced pattern of the wing venation is clearly different in Zoraptera and Embioptera, and distinct differences are also present in the zorapteran and isopteran wings (M. Engel, personal communication). The shortened tarsi (2 or 3 tarsomeres), cerci (1 or 2 cercomeres) and the vestigial ovipositor are the result of unspecific reductions (e.g. 2 or 3 tarsomeres and completely reduced cerci in Acercaria, see below). The enlargement of the femora in both groups is rather a tendency than a well-defined synapomorphic character state. As the femora are modestly sized in the specimens of Zorotypus hubbardi, examined during this study, it is not entirely clear whether this is a groundplan character state of Zoraptera. However, this presumably derived condition is distinct in putative basal fossil zorapterans such as † Xenozorotypus and † Octozoros (Engel 2003b). Head structures of Embioptera and Zoraptera differ considerably. The head of Embioptera is distinctly prognathous and ocelli are primarily absent. Unusual muscles occurring in Embioptera (tentorial retractor of the salivarium and hypopharynx, transverse muscle of the hypopharynx, suspensor muscle of tentorium; Rähle 1970) are absent in Zoraptera and vice versa. The only possibly derived feature of the head shared by both groups is the presence of a distinctly delimited anteclypeus. However, this feature is also present in several groups of Acercaria (see below). A close relationship between Zoraptera and Embioptera should be considered as one option, but the presently available arguments are not sufficient.

It should be noted here that a placement of Zoraptera among the “lower Neoptera” (see Engel and Grimaldi 2002) is far from being settled. Neither the morphological (Boudreaux 1979) nor the molecular evidence (Wheeler et al. 2001) supporting a potential clade “Polyneoptera” including Zoraptera is convincing so far.

A sister-group relationship between Zoraptera and a clade comprising Acercaria and Endopterygota was obtained in a comprehensive cladistic analysis carried out by Beutel and Gorb (2001). However, it was pointed out in that study that this is only supported by very weak evidence. A sister-group relationship with Endopterygota was suggested by Rasnitsyn (1998) based on a study of thoracic features. Unfortunately, Rasnitsyn’s (1998) investigation was only based on a single wingless specimen and was not very detailed. Rasnitsyn (1998) did not present arguments in a cladistic or Hennigian sense and besides this, his case is weakened by the fact that the morphology of the wingless forms does not represent the groundplan condition of Zoraptera (see above). Engel and Grimaldi (2000) also criticised some of his putative synapomorphies, suggesting alternative interpretations.

A hypotheses, which was proposed by Hennig (1969) and others (see Kristensen 1975, 1981, 1991), is a sister-group relationship with Acercaria. The monophyly of Paraneoptera (=Zoraptera + Acercaria) is supported by the reduced number of Malpighian tubules (six in Zoraptera, four in Acercaria), the reduced number of tarsomeres (two in Zoraptera, maximum number of three in extant Acercaria) and the concentraction of the abdominal ganglia (two ganglionic masses in Zoraptera, one complex in Acercaria). All these features are reductions, and it is unclear whether the presence of less than 5 tarsomeres is a groundplan feature of Psocoptera (Hennig 1969; but see Kristensen 1975).

Another potential paraneopteran autapomorphy is the presence of an anteclypeus (Figs. 1a, 2a, 5), which is very distinctly separated from the postclypeal area (present in Thysanoptera, Psocoptera, Heteroptera; e.g. Badonnel 1934; Cope 1940; Mickoleit 1963). However, it is not clear, whether this is indeed derived and whether it belongs to the groundplan of Acercaria. Besides this, the same condition is present in Embioptera (see above). Another apomorphic condition is the very strongly developed M. clypeobuccalis. This muscle is very large and complex in Z. hubbardi (Figs. 5, 6a, 7a, b) and in all groups of Acercaria (e.g. Weber 1929, 1938; Badonnel 1934; Risler 1954; Buckup 1957; Mickoleit 1963). Even though this is rather a gradual modification, and relatively strong and complex clypeobuccal muscles do also occur in other groups (e.g. Blattodea; Dorsey 1943; Moulins 1971), this should be considered as a potential autapomorphy of Paraneoptera. The muscle is much smaller and less complex in most other groups of insects.

Another feature shared by Zoraptera and acercarian groups is a prementum medially divided by a rim and corresponding internal ridge (see Figs. 1b, 2b, 7a). This condition is also present in Psocoptera (Badonnet 1934, ‘gouttière médiane’: Fig. 33) and Mallophaga (Myrsidea; Buckup 1957; Haub 1967). It is quite possible that this is a groundplan feature of Acercaria. The absence in acercarian subgroups may be related to strong modifications of the labium in taxa with a piercing–sucking feeding apparatus (e.g. Mickoleit 1963; Tröster 1990).

A distinct modification of the lacinia is clearly in support of Paraneoptera. Whereas the lacinia is almost always moderately broad and equipped with setae and spines along its mesal edge (e.g. Blattodea, Mantodea, Embioptera; Rähle 1970; R. G. Beutel, personal observations), the lacinia is slender in Zorotypus hubbardi (Fig. 1b), and the mesal edge is smooth and completely devoid of spines or setae (Fig. 4b). This type of lacinia is probably not very suitable as a grasping device as this is normally the case in insects with biting mouthparts. The similar laciniae of Psocoptera were referred to as “organes absolument inutiles” by Badonnel (1934). This interpretation is probably incorrect, as was pointed out by Kéler (1966). It is likely that they are used to loosen the food substrate in Psocoptera (and in Zoraptera). The zorapteran galeae, with their rake-like apical parts are probably used to collect the material and move it into the preoral cavity (corrosium).

It appears plausible to assume that a possible common ancestor of Zoraptera and Acercaria was of small size and microphagous (fungal hyphae, spores and pollen; see Gurney 1938; Riegel 1963), with chisel-like laciniae and mandibles with strongly developed grinding molae. The slender lacinia was detached from the stipes in the ancestor of Acercaria. A unique complex of cibarial sclerotizations (ovoid sclerites, filamentous ducts, mortar-and-pistill apparatus) has then evolved in Psocodea (groundplan: Psocoptera and Ischnocera; Badonnel 1934; v. Kéler 1966; Risler 1954), and piercing–sucking mouthparts in Anoplura, Hemiptera and Thysanoptera.

Even though the results of the present study seem to support the monophyly of a clade Paraneoptera, the placement of Zoraptera is not yet settled. A reliable clarification will require a cladistic analyis of an extensive and solid morphological character set and of DNA sequences of different genes. A detailed study of thoracic structures of alate forms should have high priority along with investigations of the abdomen.