New records of immature aquatic Diptera from the Foulden Maar Fossil-Lagerstätte, New Zealand, and their biogeographic implications

Studied material

The 30 immature individuals of flies described herein (Figs. 1–14) were collected in mining pit A (45.5269°S, 170.2191°E) at Foulden Maar (see Kaulfuss, 2017, Fig. 1C), which exposes a ca. 10 m thick succession of fossiliferous diatomite representing a depositional period of ca. 18,000 years. The site is registered as I43/f8503 in the New Zealand Fossil Record File (https://fred.org.nz/). Specimens were found randomly distributed throughout the stratigraphic section (no mass mortality layers were observed) and are always preserved on light-coloured diatomaceous sublaminae, which resulted from diatom blooms in warm seasons. No specimens were found in the dark, organic-rich sublaminae deposited in the cooler seasons.

All individuals are pale to dark brown, strongly compressed (compacted) specimens, mostly preserved as part and counterpart, although often only the part exhibits useful morphological details, whereas the counterpart is a faint outline or impression only. One larva specimen, mounted on a glass slide (Fig. 7), was recovered from the carbonaceous residue of the diatomite. To this end, the silica component was dissolved in a plastic beaker with water and hydrofluoric acid, the residue washed and sieved, and the larva mounted in suspension onto a coverslip with a pipette. After removal of the water, the coverslip with the specimen was permanently mounted on a slide with an epoxide-based mounting medium (Petropoxy 154) (N Butterfield, 2022, personal communication). All specimens are stored in the Museum of the Geology Department, University of Otago (OU); identifiers provided below consist of an OU collection number followed by an original field number in brackets.

Imaging

Specimens were photographed with a Canon T3 camera attached to a Nikon SMZ1000 stereomicroscope. Wetting the specimens in ethanol improved the contrast between specimens and the diatomite matrix. The single slide-mounted specimen was imaged using a Keyence BZ-9000 fluorescence microscope with either 4×, 20×, 40× or 100× objectives. We have conducted observations using an emitted wavelength of 532 nm since it was the most compatible with the fluorescence capacities of the fossil specimen (Haug et al., 2011). Stacks of images were digitally computed to single in-focus images using CombineZP (GNU) or Photoshop CS5.1 software (Adobe Systems Inc., San Jose, CA, USA).

Morphological terminology and identification

In the course of our work we normally do not use Linnean ranks (‘rankless taxonomy’) (Baranov, Schädel & Haug, 2019; Baranov et al., 2022b; Haug et al., 2020). Ranks (or “categories” sensu Mayr, 1942, p. 102) are arbitrary constructs of the human-imposed structure that does not hold ‘comparative values’ (Mayr, 1942, p. 291, line 3). In our view such arbituary constructs do not contribute to facilitation of the understanding of phylogenetic relationships between the organisms, including both species and higher phylogenetic groups, instead we are using rankles hierarchy of the monophyletic groups (Haug et al., 2020).

The morphological terminology is based on Sæther (1980), Marshall et al. (2017) and Borkent & Sinclair (2017), and specifically follows Borkent (2012) for culicomorphan pupae anatomy. In this article we describe morphotypes, i.e., distinct morphological groups of organisms. Members of each morphotype are here assumed to represent the same species, although this is often impossible to ascertain for the fossils. Most of the fossils dealt with herein are pupal exuviae (integument left after the eclosion of the adult fly) but for convenience we treat all pupae fossils and their integuments as “pupae”.

Specimens were identified using the keys provided by Wiederholm (1986), Langton (1991), Sæther (1970, 1997), Roback (1971), Forsyth (1971), Cook (1956), Colless (1983) and Winterbourn, Gregson & Dolphin (1989).

Systematic paleontology

DIPTERA Linnaeus, 1758

CHIRONOMIDAE Newman, 1838

TANYPODINAE Thienemann & Zavřel, 1916

Material: specimens OU46609 (191), OU46626 (208), OU45541 (137), OU44933 (28), OU44930 (23), OU46641 (223) (Figs. 1–3).

Pupa. Habitus. Medium-sized. All specimens preserved in dorso-ventral aspect, thus precluding detailed observation of head and legs. Body length 4.5–5.4 mm (n = 2); abdomen length 2.9–3.7 mm (n = 2); length of thorax 1.6–1.7 mm (n = 2), body differentiated into presumably 20 segments, ocular segment plus 19 post-ocular segments (Figs. 2A, 2B); anterior part of body composed of head and thorax, visible as single semi-circular structure; thorax with wings and ambulatory appendages (legs) (Figs. 1A–1F, 2A, 2B); ocular segment and post-ocular segments 1–5 (presumably) forming a distinct capsule (head capsule); mouthparts located ventrally and thus not available for observation (Figs. 1A–1F). Pupal cuticle of the head with prominent frontal apotome (frontal protrusion of the head), apparently bearing a pair of strong frontal setae (only right one is visible), sitting on the short tubercle (Fig. 3A).

Thoracic segments forming a single semi-globose structure closely enveloping the head of the pupa. Prothorax bears thoracic horns (modified first spiracle) (Figs. 2A, 2B, 3A, 3C, 3D), these mostly cylindrical, 4.7 times as long as wide, with a total length of 400 µm (n = 1, only one horn was preserved in a good condition) (Figs. 3C, 3D). Widest point of thoracic horn at base of plastron plate (surface for retention of the air film, providing a gas exchange interface), overall shape of the horn tapers slightly towards the base (Figs. 3C, 3D). Plastron and atrium connection area light colored. Plastron plate ovoid, 90 µm long and almost 90 µm wide. (Figs. 3C, 3D). Thoracic horn with apical papilla (Figs. 3A, 3C). Thorax is very wrinkly, bearing no distinct sclerotized protrusions (i.e., thoracic combs, sensu Saether, 1980). Wings and their cuticular sheath barely visible, hidden under the body, only points of the wing attachment to the mesothorax discernible (Figs. 2A, 2B).

Abdomen (posterior trunk). Abdominal cuticle extremely wrinkly. First unit of abdomen with very strongly pigmented scar (Figs. 1A–1F, 2A, 2B). Setae of abdominal units not preserved, but sturdy tecae of some of them can be seen on some abdominal tergites. Abdominal tergites 3 and 4 bearing tecae of five dorsal setae each, abdominal tergite 5 with at least one pair of tecae of dorsal setae (Fig. 2A). Traces of lateral setae on abdominal units 2–6 not discernible. Abdominal unit 8 with traces of bases of some lateral setae. Trunk end (abdominal unit 9 plus remnants of abdominal unit 10) bears anal lobes (paddles) (Fig. 2A), these semi-circular, broadly rounded, 1.6 times longer than wide. Anal lobes bearing a fringe of spine-like, short setae, present in the distal part of the lobes only (Figs. 2A, 2B, 3D).

Taxonomic attribution. These specimens are representatives of Chironomidae based on the following combination of pupal characters: thoracic horn with strong plastron plate; strongly sclerotized arches from the anterior parts of the abdominal tergites; terminus of trunk without articulated terminal paddles (Figs. 2A, 3D). Within Chironomidae, this morphotype falls within the Tanypodinae based on the following combination of characters: thoracic horn well developed, with prominent plastron; anal lobes rounded, with fringe of the setae along the outer edge of the anal lobes, anal lobes completely confluent and with overlapping inner edges (Fig. 3A) (Fittkau & Murray, 1986). It is difficult to place this morphotype within Tanypodinae, due to many characters being un-observable. However, the structure of the thoracic horns, in particular light colored connection between plastron and atrium and presence of apical papilla and overall shape (broadly rounded) of the anal lobes are pointing to the Procladiini or Anatopyiini, with Procladius being a particularly close match, due to spine like setae present only on the distal part of the anal lobes (Fittkau & Murray, 1986).

Remarks. The are no extant representatives of Tanypodinae with morphologically similar pupae in New Zealand, but Forsyth (1971) reported a number of species of “Anatopynia”, of which some share a certain degree of similarity with this fossil morphotype. In particular, “Anatopynia” anatarctica (Hudson, 1892) shares with the new fossil morphotype the broad, rounded shape of the anal lobes and the general structure of the thoracic horn (Forsyth, 1971: Fig. 2). However, “A.” antarctica has prominent distal protruding points on the anal lobes, which are absent from the new morphotype from Foulden Maar. Additionally, the inner edges of the anal lobes are overlapping in the new morphotype (Fig. 3A), which is characteristic for representatives of Coelotanypus, but not for any of the representatives of “Anatopynia” recorded from New Zealand (Forsyth, 1971; Fittkau & Murray, 1986).

CHIRONOMINAE Macquart, 1838

CHIRONOMINI Macquart, 1838

cf. Chironomus Meigen, 1803

morphotype 1 (pupae)

Material: OU47058 (252), OU46654 (236), OU46645 (227), OU46627 (209), OU46625 (207), OU46620 (202), OU45553 (149), OU45543 (139), OU47051 (245) (Figs. 4–6).

Pupa. Habitus. Medium-sized, coma-shaped (in lateral aspect). Most specimens preserved in dorso-ventral aspect, thus precluding detailed observation of head and legs. Body length 4.7–6.2 mm (mean = 5.6 mm, sd = 430 µm) (n = 8); abdomen length 3.6–4.5 mm (mean = 4.2 mm, sd = 355 µm) (n = 8); length of thorax 1.3–2.3 mm (mean = 1.6 mm, sd = 330 µm) (n = 8). Body differentiated into presumably 20 segments, ocular segment plus 19 post-ocular segments (Figs. 4A–4H, 5A–5H, 6A–6E); anterior part of body composed of head and thorax, visible as single semi-circular structure; thorax bears wings and ambulatory appendages (legs) (Figs. 6A, 6B); ocular segment and post-ocular segments 1–5 (presumably) forming a distinct capsule (head capsule); mouthparts located ventrally and thus not discernible (Figs. 5A–5H, 6A, 6B). Pupal cuticle of head with prominent frontal apotome (frontal protrusion of the head). Frontal apotome bears a pair of strongly curved, conical cephalic tubercles with a pair of strong frontal setae (Figs. 6A–6C).

Thorax with three pairs of ambulatory appendages (fore-, mid- and hindlegs) on the pro-, meso-, and metathorax, respectively. Thoracic segments forming a single semi-globose structure closely enveloping the head of the pupa. Anterolateral and anteromedian tubercles absent. Forelegs folded around dorsal side of wing (Figs. 6A, 6B). No traces of the thoracic horns were found, but specimen OU45543 (139) shows the presence of the tracheal scar (place where trachea is passing through the thorax cuticle into the thoracic horns). Mesothorax with a pair of wings and a pair of ambulatory appendages (midlegs). Midlegs situated medially to forelegs, looping around wing, distal part of the loop lying on the abdomen, beyond the distal end of the wing. (Figs. 6A, 6B). Metathorax with a pair of ambulatory appendages (hindlegs); halteres not discernible. Hindlegs almost entirely hidden behind wings (Figs. 6A, 6B).

Abdomen (posterior trunk). Made up of nine visible abdominal units. Tergal armament most complete and best preserved in specimens OU46654 (236), OU47051 (245) and OU47058 (252) (Figs. 5B, 5G, 5H, 6A, 6B). Abdominal tergites unmodified, without plates or spine mounds. Abdominal tergite 1 bare. Abdominal tergite 2 with fine shagreen pattern and continuous row of hooks at posterior edge (Figs. 6A, 6B). Abdominal tergites 3–5 with uniform shagreen and strong oval patch of longer, dark spines located on the median line of tergite, touching the posterior edge of the segment. Tergite 2 with hook row narrowly interrupted or continuous, but definitely without a wide gap. Tergites 2–4 without anterior band of shagreen. Tergite 6 apparently mostly bare, without visible shagreen, bearing similar medio-posterior patch of strong, dark spines (Figs. 6A, 6B). Tergites 7 and 8 mostly bare, without visible shagreen (Figs. 6A, 6B). Abdominal unit 8 bearing two strong anal combs postero-lateraly (Figs. 6D, 6E), these made up of four strong spines, the outermost being the longest and the others getting shorter towards the innermost spine. Anal lobes semi-circular, with strong fringe of at least 50 setae (difficult to count) on each lobe (Fig. 6D), without any processes or modified setae. Genital sacs of the male without spines. No other setae are preserved on the abdomen of this morphotype. In specimen OU46645 (227), a part of a male hypopigium is visible through the cuticle of the genital sack. Strong, blunt anal point visible, alongside a long, curving gonostyle, joined to a gonocoxite (Fig. 5D).

Taxonomic attribution. Specimens of this morphotype are representatives of Chironomidae based on the following combination of pupal characters: strongly sclerotized arches from the anterior parts of the abdominal tergites; terminus of trunk without articulated terminal paddles (Figs. 5A–5H, 6A, 6D). Within Chironomidae, this morphotype can be interpreted as an ingroup of Chironominae because of the diagnostic combination of characters: abdominal tergite 8 with a strong anal comb, anal lobes with a well-developed fringe of setae, gonostylus and gonocoxite conjoined rigidly, no articulation visible (Fig. 5D) (Wiederholm, 1989). This morphotype falls within Chironomini based on the following combination of characters: tergites 2–6 without paired patches of shagreen, bearing rectangular fields of shagreen instead, wing sheath without “nose” protrusion on the distal end (Wiederholm, 1989). Within Chironomini, this morphotype appears to be a representative of Chironomus based on the following combination of characters: cephalic tubercles present but without apical field of spines, frontal warts absent, frontal setae present, anterolateral and anteromedian tubercles absent; tergites of the abdomen bearing large but unmodified fields of shagreen, tergites 2–4 without anterior band of shagreen; hook row of tergite 2 without a wide gap, abdominal tergites without anteromedian toothed lobe, abdominal tergite 6 without posteromedian mound of spines, anal lobes without brushes of dark setae and without long spines; anal lobe bearing uniform fringe of setae, forked process, genital sacs of males without spines, anal comb in form of elongated spine with three points (Wiederholm, 1989, Forsyth & Martin, 2014).

CHIRONOMINI Macquart, 1838

cf. Chironomus Meigen, 1803

morphotype 2 (larva)

Material: OU47488 (268) (Fig. 7).

Larva. Habitus. A single specimen mounted on a glass slide (Figs. 7A–7D). Larva with well-preserved body and head capsule cuticle, with many microscopic details such as structure of submentum or mandible apparent. Head capsule well developed, with complete sclerotization and overall non-retractable. Ocular segment and post-ocular segments 1–5 (presumably) forming a distinct capsule (head capsule). Head capsule without conspicuous labral fans and well-developed epipharyngeal complex (of which only premandibles are visible). Head capsule bears pair of mandibles. Mandibles with pronounced apical tooth and three internal teeth (Fig. 7C, 7D). Premandibles apparently with more than three apical teeth. Mentum (part of labium) well pronounced, with seven pairs of lateral teeth, with first part of lateral teeth about four times shorter than second pair, with wide gap between them.

Head and thorax not conjoined together; no suction discs at the abdomen; abdominal cuticle well preserved, but without visible setae; respiratory system lacking developed trachea or external spiracles (apneustic type); thorax segments well distinguishable; prothorax and abdominal unit 9 with paired parapods; abdominal units 1 and 2 without parapods; group of strong, downward pointing preanal setae absent on the trunk end.

Taxonomic attribution. This morphotype falls within Chironomidae based on the combination of the following characters: larva with well-developed, complete and non-retractable head capsule; head capsule without conspicuous labral fans and well-developed epipharyngeal complex; head and thorax not conjoined together; no suction discs at the abdomen; respiratory system lacking developed trachea or external spiracles (apneustic type); thorax segments well distinguishable; prothorax and abdominal unit 9 with paired parapods; abdominal units 1 and 2 without parapods; group of strong, downward pointing preanal setae absent on the trunk end (Ekrem et al., 2018). Within Chironomidae, this larva falls within Chironominae based on the following combination of characters: antenna not retractable, non-annulate, labrum without row of the overlapping lamellae, premandible present, submentum with 15 teeth in three distinct groups, symmetrically distributed on submentum (Wiederholm, 1983; Cranston, 2019).

The preservation of the head capsule is not conducive for further identification of the specimen but the general habitus is highly reminiscent of that of Chironomus Meigen, currently represented in New Zealand by at least six species (Boothroyd & Forsyth, 2011). We propose affinity of this morphotype to Chironomus based on the following combination of the characters: mandible with three inner teeth, premandible with seemingly more than three teeth, central tooth of the mentum trifid (Wiederholm, 1983; Cranston, 2019). It is important to note that with the characters available it is impossible to positively differentiate this morphotype from Goeldichironomus Fittkau, 1965.

cf. CHIRONOMINAE Macquart, 1838

Chironomidae morphotype 1

Material: OU45549 (145), OU46608 (190) (Fig. 8).

Pupa. Both specimens are too poorly preserved for a detailed description (Figs. 8A–8D).

Taxonomic attribution. These two specimens of pupae belong to a separate morphotype, which is difficult to place due to the missing characters of the distal end of the abdomen, yet their overall habitus is highly reminiscent of that of Chironomidae. Defining feature of this morphotype is a strong, protruding spine on the posterior edge of tergite 2 (Figs. 8A–8D). Since value of leg ratio (LR = length of tarsomere one divided by the length of tibia) is >1, it is probable that this morphotype belongs to Chironomidae (Wiederholm, 1989).

CHAOBORIDAE Edwards, 1912

Chaoboridae morphotype 1

Material: Specimens OU47487 (163), OU46642 (224), OU46631 (213), OU46651 (233), OU46653 (235) (Figs. 9–13).

Pupa. Habitus. Medium-sized, coma-shaped (in lateral aspect). Body length 5.2–6.0 mm (n = 3, mean = 5.5 mm, sd = 490 µm); abdomen length 3.8–4.3 mm (n = 3, mean = 4.0 mm, sd = 320 µm); length of thorax 1.3–18.8 mm (n = 4, mean = 1.5 mm, sd = 220 µm). Body differentiated into presumably 20 segments, ocular segment plus 19 post-ocular segments (Figs. 13A, 13B); anterior part of body composed of head and thorax, visible as a single globose structure; thorax bears wings and ambulatory appendages (legs) (Figs. 11A, 11B, 12A, 12B, 13A, 13B); mouthparts located ventrally and short, ending before attachment of first ambulatory appendages (forelegs) (Figs. 11A, 11B, 12A, 12B, 13A, 13B). Ocular segment recognizable by its appendage derivative, clypeo-labral complex, and a pair of large compound eyes. Labrum and clypeus present, but their shape obscured by deformation and preservation in lateral aspect (Figs. 11A, 11B, 12A, 12B). Antennae curved around the head, ending beneath the head at about mid-length to 0.8 of the length of the wings. Antennae attached to massive, rounded pedicellus (antennal element 2) (Figs. 11A, 11B, 12A, 12B). Maxilla recognizable by maxillary palpus, palpi poorly preserved. Post-ocular segment 5 recognizable by its appendages, forming the labium [conjoined left and right maxillae]. Labium mostly obscured in all specimens, with no details visible (Figs. 11A, 11B, 12A, 12B). Thorax bears three pairs of ambulatory appendages (fore-, mid- and hindlegs) on the pro-, meso-, and metathorax, respectively. Thoracic segments forming a single semi-globose structure, closely enveloping the head of the pupa. Ambulatory appendages of the thorax folded around and under the wings (Figs. 11A, 11B, 12A, 12B). Prothorax bears thoracic horns (modified first spiracle). Thoracic horns (respiratory organs) absent in all specimens. Prothorax bears first thoracic appendages (forelegs). Forelegs running posteriorly, upwards anteriorly to upper edge of eye and then downward to the apical edge of wing (Figs. 11A, 11B, 12A, 12B, 13A, 13B). Mesothorax bears a pair of wings and a pair of ambulatory appendages (midlegs). Midlegs situated medially to foreleg, looping around the wing, distal part of the loop lying on the abdomen, beyond the distal end of the wing. Distal parts of midlegs loop again under the wing (Figs. 11A, 11B, 12A, 12B, 13A, 13B). Hindlegs almost entirely hidden behind coxae of the fore- and midlegs and wings (Figs. 11A, 11B, 12A, 12B, 13A, 13B).

Abdomen (posterior trunk). Abdominal units 1–8 with setae on the pharate adult tergites visible through the pupal cuticle. Setae radiating from the median line of the abdomen diagonally, so as to form pointed bundleds of setae at the posterio-lateral part of tergites 1–8 (Figs. 11A, 11B, 13A, 13B). Trunk end (abdominal unit 9 plus remnants of abdominal unit 10) bears hypopigium (male genitalia) (only visible on specimen OU46653 (235)). Hypopigium consists of paired gonocoxites (II) of abdominal unit 9 and paired gonostyli articulated at the distal end of gonocoxites (Figs. 13C, 13D). Gonocoxites ca. 250 µm long, gonostyli ca. 430 µm. Gonostyli straight, ending with short, rounded apical setae (megaseta) (Figs. 13C, 13D). Gonocoxites densly covered with strong setae. No traces of anal lobes present on the trunk end (Figs. 13C, 13D).

Taxonomic attribution. We interpret this new morphotype as an ingroup of Chaoboridae based on the specific combination of the following characters: Pupa: mouthparts short, not reaching beyond coxae of anterior legs; bundles of diagonally oriented setae on tergites of the abdominal units (the latter character is a autapomorphy of Chaoboridae) (Figs. 11A, 11B, 13A, 13B) (Borkent & Grimaldi, 2004; Borkent, 2012). Some characters of the adult male were available for examination through the cuticle of the pharate adult (inside the pupa) in specimen OU46653 (235) (Figs. 13C, 13D). However, the poor preservation does not allow for a closer taxonomic attribution as characters of the adult legs and wings are not discernable in this specimen. All the available specimens of Chaoboridae from Foulden Maar are missing anal lobes and thoracic horns. These characters are crucial for diagnosis and taxonomic attribution of representatives of Chaoboridae (Cook, 1956; Sæther, 1970). Therefore, we cannot attribute this new morphotype to any ingroup of Chaoboridae, although the general shape of the pupa and shape of the hypopigium are very similar to pupae of Chaoborus Lichtenstein (Cook, 1956; Sæther, 1970, 1997; Borkent, 2012). We thus suggest that the Chaoboridae morphotype from Foulden Maar is either a representative of Chaoborus or closely related to it. To further validate this assumption, we will require additional material with preserved anal lobes and thoracic horns. The loss of anal paddles in pupae from the finely laminated Foulden Maar diatomite is likely due to the fragility of the “paddle” attachments (anal lobe). A similar preservation is present at the Eocene Kishenehn Formation, USA, and the Miocene McGrath Flats Formation, Australia, where otherwise exquisitely preserved Chaoboridae pupae are frequently missing their anal “paddles” (anal lobes) (Baranov et al., 2022b; McCurry et al., 2022).

BRACHYCERA (?)

Material: OU44944 (43), OU44981 (90), OU44982 (91), OU44996 (105), OU45559 (155), OU46644 (226), OU46652 (234), OU46655 (237) (Fig. 14).

The examined material contains eight specimens, seemingly immatures of a holometabolan; all share similar dark-brown or reddish-brown colours (Figs. 14A–14H). All specimens have at least 11 body units and some have their supposed tergites split in half, giving an impression of the post-eclosion pupal exuvia (Figs. 14C, 14D, 14H). Tergites are split in the middle. Some specimens have strong setae on the edges of the tergites (Figs. 14B, 14D, 14F). Specimens OU44944 (43) and OU44996 (105) are more of a spindle-shape and have of an overall appearance of dipteran larvae, rather than pupal exuvia or puparia (Figs. 14A, 14G). Unfortunately, there are no diagnostic characters allowing for a closer taxonomic placement of these specimens. Based on the habitus similarity, we hypothesize that these specimens represent larvae and puparia (larval last stage exuvia, covering a pupa of brachyceran flies) of brachyceran flies, likely representatives of Cyclorrhapha (i.e., see Ferrar, 1987, vol 2 Fig. 6.155 (p. 568)).

Paleontological significance and paleoecology

Fossil deposits in southern New Zealand document the presence of various lentic habitats and associated freshwater faunas during the early and middle Miocene (e.g., Douglas, 1986; Pole, Douglas & Mason, 2003; Lee et al., 2016; Kaulfuss et al., 2018a, 2020). Previously reported fossils of the groups Odonata, Megaloptera, Coleoptera, Diptera and Trichoptera provide a glimpse of the aquatic insect fauna in these paleo-habitats (Kaulfuss et al., 2015, 2018a, 2018b; Schmidt et al., 2018; Baranov et al., 2022a), but the diversity, biogeography and ecological role of individual groups remain to be explored in detail. Our study gives insights into the aquatic dipteran fauna in a small, isolated maar lake in southern New Zealand in the earliest Miocene (23 mya), shortly after maximum marine inundation of most land area at ~25 mya. The larvae and pupae described here from the Foulden Maar diatomite indicate the presence of a dipteran fauna consistent in its taxonomic richness with the diversity of merolimnic flies recorded from other Paleogene and Neogene lacustrine deposits: Kishenehn Formation, Eocene, USA, eight morphotypes (Baranov et al., 2022b), McGrath Flats, Miocene, Australia, five morphotypes (McCurry et al., 2022), Messel, Eocene, Germany, eight morphotypes (Paleobiology Database, 2023), Randecker Maar, Miocene, Germany, one morphotype (Paleobiology Database, 2023).

The non-biting midges (Chironomidae) from Foulden Maar include Tanypodinae, one pupal and one larval morphotype in Chironomini (Chironominae), both resembling Chironomus, and a further chironomid morphotype of uncertain identity, possibly in Chironominae. The only non-biting midge fossils previously reported from New Zealand are four adult specimens of Bryophaenocladius Thienemann (Orthocladiinae) from late Oligocene amber (Schmidt et al., 2018). Phantom midges (Chaoboridae) from Foulden Maar are represented by a pupal morphotype congeneric or closely related to the widespread and speciose extant group Chaoborus. These pupae are the first fossil record of phantom midges from New Zealand. Intriguingly, phantom midges are absent in the extant fauna of New Zealand (see below). Adult life stages of non-biting and phantom midges have not yet been found in the Foulden Maar diatomite. The presence of immature aquatic brachycerans in the Foulden Maar paleo-lake is documented by eight larvae and puparia of uncertain systematic position. Rare isolated wings of adult brachyceran flies (either representatives of Muscidae or Acalyptrata) have previously been reported from the fossil site (Kaulfuss et al., 2015), but affiliation of these with any of the immature aquatic specimens cannot be established due to incomplete preservation.

The comparatively small sample of identifiable insects (n = 253) from Foulden Maar includes a relatively high proportion of immature aquatic dipterans (16%), suggesting that these life stages were a common component in this limnic paleo-ecosystem. Of these, non-biting pupae are most common (63% of immature dipterans), followed by the brachyceran-type (20%) and by phantom midges (17%). Together with other Crustacean forms, they likely provided a food source for fish, in particular for Galaxias effusus, which is commonly found as larvae, juvenile and adult specimens in the diatomite (Lee, McDowall & Lindqvist, 2007; Kaulfuss et al., 2020). The most common type of coprolite at Foulden Maar is most likely derived from Galaxias effusus and consists of mineral grains, plant material and common euarthropodan fragments (Kaulfuss, 2013). The latter are yet to be studied in detail and it is currently not known if non-biting or phantom midge remains are present in these coprolites. In any case, the occurrence of euarthropodan remains in these coprolites conforms to a diet of aquatic and terrestrial insects observed in most extant Galaxias species in New Zealand (McDowall, 2010). Immature dipterans are also commonly preyed on by various groups of aquatic eurthropodan animals such as water mites, dragonflies and damselflies larvae, aquatic bugs and beetles (e.g., Armitage, Cranston & Pinder, 1995; Martin & Gerecke, 2009; Ferrington, 2008; Klecka & Boukal, 2012). Although this probably was the case in the Foulden Maar lake, the available fossil data are insufficient for establishing specific predator-prey relations.

Assuming the ecology of the Chaoboridae midges from Foulden Maar concurs with that of extant relatives (e.g., Macdonald, 1956; Rudstam, 2009; Hare & Carter, 1986), its larvae may have been abundant in the pelagic and littoral zones of the maar lake, feeding on small eucrustaceans (e.g., of the groups Copepoda and Cladocera), benthic organisms and dipteran and other insect larvae, and possibly also ingesting readily available phytoplankton. As in extant species of Chaoborus, late stage larvae (3^rd and 4^th instar), probably exhibited diurnal vertical migration, preying in the epilimnion at night and migrating into deeper, oxygen-depleted zones of the monimolimnion or sediment to avoid fish predators during day (Macdonald, 1956; Hare & Carter, 1986).

Chironominae morphotype 2 is closely resembling larvae of the extant species of Chironomus and probably had a similar ecology, inhabiting soft sediments and relying on acquisition of food by bioirrigation, pumping water containing organic particles through their burrows in the sediment (Hamburger, Dall & Lindegaard, 1994). Tanypodinae is represented at Foulden Maar by specimens of probable Procladiini. Extant representatives inhabit fine-grained sediments, where they rely heavily on bioirrigation for food acquisition (Boesel, 1974; Matisoff & Wang, 1998). The larvae also prey on other bioirrigating animals, such as worms of the group Tubificidae (Soster & McCall, 1989) and are capable of utilising other sources such as suspended organic particles and detritus (Boesel, 1974; Matisoff & Wang, 1998).

The immature dipterans at Foulden Maar are primarily exuviae of pupae, left after eclosion of the adult, terrestrial stage. Pupae of both non-biting and phantom midges are short-lived, lasting only for several hours to several days, but are ecologically important in providing large amounts of organic matter in pulses to the higher order consumers in lakes and rivers (Lehman et al., 1998; Wagner, Volkmann & Dettinger-Klemm, 2012).

Biogeographical implications

Taphonomy

Immatures of the groups Chironomidae and Chaoboridae may be common in Mesozoic and Cenozoic lacustrine settings, concurring with their aquatic lifestyle (e.g., Sinichenkova & Zherikhin, 1996; Johnston & Borkent, 1998; Baranov et al., 2022b; McCurry et al., 2022). However, in Cenozoic maar lakes with well-documented insect faunas (>4,000 studied specimens) immature aquatic midges (and other groups of aquatic insects) are typically absent or only present in small numbers. This has been attributed to supposed unfavourable limnic conditions in these small but deep lakes or to taphonomic biases, which favour the preservation of larger and more compact adult insects in the profundal sediments that are usually excavated for fossils (Lutz, 1991, 1997; Wedmann, 2000; Wedmann, Poschmann & Hörnschemeyer, 2010; Wedmann et al., 2018; Wappler, 2003). At the Eocene Eckfeld Maar, for instance, the study of ~4,600 insects yielded only two pupae and several larval cases of Chironomidae, and no chaoborid midges were found (Wappler, 2003). No aquatic stages of non-biting or phantom midges were reported from the rich insect faunas of the Oligocene Enspel Maar (Wedmann, Poschmann & Hörnschemeyer, 2010) and the Paleocene maar of Menat (Wedmann et al., 2018). Immature non-biting and phantom midge body fossils are also absent in the ‘oilshales’ of the Messel Maar, although their remains are frequently encountered in fish coprolites, indicating ecologically important populations in this Eocene maar lake (Richter & Baszio, 2001; Richter & Wedmann, 2005; Wedmann & Richter, 2007). Exceptions appear to be the Miocene maar lakes at Öhningen and Randeck, where aquatic insects including immature Diptera have been reported as being relatively common (Heer, 1865; Joachim, 2010).

At Foulden Maar, the proportion of immature aquatic dipterans at is comparatively high, which may reflect preferential conditions for the preservation of small aquatic insects in this Miocene lake. Several taphonomic studies have highlighted the role of diatom mats in the exceptional preservation of fossil biotas (Harding & Chant, 2000; O’Brien, Meyer & Harding, 2008; Iniesto et al., 2016; Olcott et al., 2022). Diatomaceous microbial mats may facilitate fossilisation by entrapping and transporting macrobiota through the water column quickly and by stabilising the sediment surface on the sea/lake floor (Harding & Chant, 2000; Olcott et al., 2022). Additionally, extracellular polymeric substances secreted by diatomaceous mats may form a chemical microenvironment (microbial sarcophagus) that enables fossilisation by delaying decay and inducing biomineralization of euarthropods, vertebrates and plants (Iniesto et al., 2016; O’Brien, Meyer & Harding, 2008; Olcott et al., 2022). At Foulden Maar, all euarthropodan fossils are preserved in laterally continuous, light-coloured laminae essentially composed of siliceous diatom frustules. The dominating species is Encyonema jordaniforme Krammer, 1997, a pennate and likely mucilaginous diatom that flourished in the upper water column of lake Foulden and formed annual diatom blooms over the lake’s estimated life span of 130,000 years (Harper et al., 2019). Centric diatoms are present as minor constituents in the lake sediment (Kaulfuss, 2017). Although taphonomic processes for the Foulden Maar biota are yet to be studied in detail, it is likely that diatomaceous mats essentially composed of E. jordaniforme might have provided a taphonomic pathway for the preservation of small aquatic larvae/pupae and other macrobiota in the diatomite.

Sediment colour might also have an impact on the apparent proportion of aquatic dipterans and other small euarthropods in fossil deposits. The Cenozoic maars at Eckfeld, Messel, Enspel and Menat, where immature non-biting or phantom midges are absent or rare, are primarily made of dark, organic-rich clay and mudstones, which makes spotting small fossils of similarly dark colour difficult. For Messel Maar, Wedmann & Richter (2007) argued that phantom midge larvae are likely present in the sediment, but their weakly sclerotised, translucent body cannot be seen in the organic shales. At Foulden Maar, the light-coloured (white to beige) diatomaceous laminae exhibit a pronounced colour contrast to embedded, typically brown or black fossil organisms. This contrast likely facilitates spotting of small euarthropods such as dipteran pupae in the field, and it might be one factor for the relatively high abundance of aquatic dipterans relative to most other maar-type Lagerstätten. The other two other Cenozoic maar lakes with a relatively high abundance of immature aquatic Diptera also consist of, or at least include light-coloured lithologies. At the mid-Miocene Öhningen Maar, many insects were recovered from white limestones (“Weißer Schieferstein”) and light-grey marlstones (“Kesselstein”) (Rasser et al., 2023). Similarly, the main insect-bearing lithologies with immature aquatic dipterans at the mid-Miocene Randecker Maar appear to be calcareous and marly laminites and limestones of lighter colour (Westphal, 1963; Joachim, 2010; Rasser et al., 2013).