Larval morphology of the avian parasitic genus Passeromyia: playing hide and seek with a parastomal bar

The enigmatic larvae of the Old World genus Passeromyia Rodhain & Villeneuve, 1915 (Diptera: Muscidae) inhabit the nests of birds as saprophages or as haematophagous agents of myiasis among nestlings. Using light microscopy, confocal laser scanning microscopy and scanning electron microscopy, we provide the first morphological descriptions of the first, second and third instar of P. longicornis (Macquart, 1851) (Diptera: Muscidae), the first and third instar of P. indecora (Walker, 1858) (Diptera: Muscidae), and we revise the larval morphology of P. heterochaeta (Villenueve, 1915) (Diptera: Muscidae) and P. steini Pont, 1970 (Diptera: Muscidae). We provide a key to the third instar of examined species (excluding P. steini and P. veitchi Bezzi, 1928 (Diptera: Muscidae)). Examination of the cephaloskeleton revealed paired rod‐like sclerites, named ‘rami’, between the lateral arms of the intermediate sclerite in the second and third instar larva. We reveal parastomal bars fused apically with the intermediate sclerite, the absence of which has so far been considered as apomorphic for second and third instar muscid larvae. Examination of additional material suggests that modified parastomal bars are not exclusive features of Passeromyia but occur widespread in the Muscidae, and rami may occur widespread in the Cyclorrhapha.


INTRODUCTION
Passeromyia Rodhain & Villeneuve, 1915 is a small genus of the fly family Muscidae, limited to the Old World (Skidmore, 1985). The genus is represented by five or six species, due to the uncertain position of P. pruinosa (Wulp, 1880) (described as Cyrtoneura pruinosa) for which the type specimens are lost (Pont, 1974). Several species have a relatively limited distribution. Passeromyia longicornis (Macquart, 1851) is endemic to Tasmania, P. indecora (Walker, 1858) is known from mainland Australia and Fiji, and P. veitchi Bezzi, 1928 occurs only in Fiji (Pont, 1974). Passeromyia steini Pont, 1970 andP. heterochaeta (Villenueve, 1915) have wider distributions; the former from Oriental and Australotropical regions and the latter from Afrotropical, Palaearctic and Oriental regions Pont, 1974). Adult flies feed on rotting plants, resin and the faeces of mammals and birds (Pont, 1974;Zumpt, 1965). However, the natural history of the immature stages is what attracts most attention from entomologists, veterinarians and conservation biologists. Larvae of Passeromyia are exclusively found in birds' nests where most species act as obligate parasites of passeriform nestlings (Skidmore, 1985) and other avian host orders (Nelson & Grzywacz, 2017;Pont, 1974). The association with birds' nests is uncommon among muscid flies (Ferrar, 1987), and Passeromyia and the New World Philornis Meinert, 1890 are the only genera containing members that are obligate parasites of nestlings (Grzywacz et al., 2015;Skidmore, 1985). Similar to Philornis, Passeromyia display various larval habits, ranging from coprophagy and saprophagy through to external and subcutaneous parasitism (Skidmore, 1985). Passeromyia steini scavenges on excreta and dead nestlings (Roberts, 1940). Passeromyia heterochaeta pierce the skin of nestlings and feed externally on blood, while P. indecora and P. longicornis burrow under the skin of the host to feed on blood or tissue (Edworthy, 2016;Skidmore, 1985). The pathogenic effects of larvalinduced myiasis in birds is well-established for Philornis with some species reported to severely impact fitness and population dynamics in New World birds (Bulgarella et al., 2015;Dudaniec et al., 2006).
Less is known about the impacts of parasitism by Passeromyia on their avian hosts, however, growth reductions and increased host mortality are among the pathogenic effects reported in early accounts (Poiani, 1993;Skidmore, 1985). More recently, surveys in Tasmania revealed an increase in morbidity and mortality among nestlings of the endangered forty-spotted pardalote (Pardalotus quadragintus Gould, 1838) associated with parasitism by P. longicornis (Edworthy, 2016;Edworthy et al., 2019).
Despite the association of Passeromyia larvae with a wide range of bird species and their impact on host fitness, knowledge of the natural history of the immature stages is fragmentary. For example, the peculiar first instar larvae of Passeromyia are well-known for their unique, long filamentous processes on the anal division, yet details of the cephaloskeleton are scant (Skidmore, 1985) despite the utility of this feature for diagnostics and interpretation of feeding habits. Existing literature concerning the larval morphology of Passeromyia encompasses two early contributions that were limited in material and scope (Skidmore, 1985;Zumpt, 1965). The well-established relationship between larval cephaloskeletal structure and evolutionary adaptation to feeding strategy (Ferrar, 1979;Roberts, 1970;Skidmore, 1985) has been utilized in taxonomic and systematic research (Grzywacz et al., 2021;Szpila, 2010). Morphological analysis of the larval stages has traditionally involved examination by light microscopy, and more recently, light microscopy combined with scanning electron microscopy (SEM) (e.g., Grzywacz, 2013;Grzywacz & Pape, 2014;Velásquez et al., 2013). Due to optical limitations in resolution, illumination and depth of field, light microscopy is largely inadequate for the precise recognition of, and interactions between, the minute sclerites in the cephaloskeleton . The application of confocal laser scanning microscopy (CLSM) overcomes these issues because it allows the visualization of fine, complex, autofluorescent larval structures . CLSM provides high resolution, highfidelity imaging and 3D reconstruction of examined structures (Szpila et al., 2021).
The objectives of this paper are to fill the knowledge gap on the larval morphology of Passeromyia to increase diagnostic capacity, and to demonstrate the advantages of applying CLSM technology in morphological studies of dipteran larvae. A combination of light microscopy, SEM and CLSM was applied to newly obtained field-collected specimens, while light microscopy was applied only to older museum material. All larval stages of P. longicornis and the first and third instars larvae of P. indecora were examined for the first time. Redescriptions of the larval morphology of P. heterochaeta and P. steini are provided.

MATERIAL AND METHODS
Larvae of P. longicornis were collected from the nests of Pardalotus quadragintus in 2017-2018 during fieldwork on Bruny Island (southern Tasmania, Australia). Passeromyia indecora were obtained from a pigeon nestling in October 2015 at Bluewater, Queensland (Australia).
For details of the identification and further information on study sites, see Edworthy (2016), Alves et al. (2020) and Nelson and Grzywacz (2017). Previously identified larval specimens of P. heterochaeta and P. steini were obtained from collections at the Natural History Museum, London, UK.
Examination by light microscopy involved destructive (P. indecora and P. longicornis) and non-destructive (P. heterochaeta and P. steini) protocols. The destructive protocol involved slidemounting the anterior body region of specimens in Hoyer's medium using concave slides. The non-destructive examination of museum collection material was preceded by the dehydration of larval specimens by serially increasing the ethanol (EtOH) concentration (80%, 90%, 99.5%) followed by immersion in methyl salicylate (Niederegger et al., 2011). Whole specimens were examined under a stereomicroscope without slide-mounting, after which the methyl salicylate was washed off and the larvae were rehydrated by decreasing the EtOH concentration (99.5%, 90%, 80%). Slidemounted specimens were examined and photographed with a Nikon 8400 digital camera mounted on a Nikon Eclipse E200 microscope (Nikon Corp., Tokyo, Japan).
SEM preparation involved dehydration of specimens at 80.0%, 90.0% and 99.5% EtOH, followed by critical point drying in carbon dioxide (CO 2 ) with an Autosamdri ® -815, Series A critical point dryer According to , material stored in Hoyer's medium is not adequate for examination by CLSM due to strong absorption of the emitted light by soft tissues and limited autofluorescence of the cephaloskeleton. Therefore, larval specimens prepared with Hoyer's medium and newly obtained ethanol-preserved material were transferred to 10% potassium hydroxide (KOH). Material was prepared according to the protocol by Szpila et al. (2021) with the following modifications: (1) adjustment of the tissue maceration interval for specimens of various sizes (16-18 h for first instar larvae to 24-50 h for second and third instars larvae) to avoid over-maceration and consequent CLSM image quality reduction; (2) the placement of larvae in a drop of viscous glycerine. Following tissue maceration in KOH, larval specimens were transferred to 80% EtOH to dehydrate for 15 min. Each sample was placed in a drop of glycerine on a cavity slide and covered with a coverslip. Sequential scanning of samples at various excitation wavelengths (488 nm, 561 nm, 633 nm) was conducted using a Leica TCS SP8 Confocal Laser Scanning Microscope To obtain the appropriate image quality and sufficient data to generate a 3D model, the number of z-steps was individually adjusted to 234, 328-391 and 433 z-frames for sequential larval stages of P. longicornis and 304 and 305-555 z-frames for the first and third larval instars of P. indecora, respectively. Specimens were deposited in the collection of the Department of Ecology and Biogeography, Nicolaus Copernicus University, Toru n, Poland. Larval terminology follows Courtney et al. (2000) with several modifications to general morphology proposed by Szpila and Pape (2005). Family-specific structures follow Skidmore's (1985) terminology with a few modifications proposed by Grzywacz (2013).
Given the numerous morphological similarities between members of the genus, descriptions of the four Passeromyia species are combined to avoid repetition, and species-specific traits are highlighted within the combined description.

First instar cephaloskeleton
The cephaloskeleton of the first instar consists of paired mouth-hooks    In the first instar, spines are distinct, light-coloured on the first thoracic segment and colourless on the remaining segments ( Figure 2a,b).

Second instar cephaloskeleton
In the second instar, the asb are composed of distinct, dark coloured spines, arranged individually or in very short rows (Figure 2d, g). In the second instar of P. steini, the asb on t1-a5 are distinct with brownish-dark spines and indistinct on a6-7. In P. longicornis, t1-a3 have dark spines, while a4-5 are colourless. On a6-7, the asb are indistinct, yet on a7, strong, individually arranged, dark-pointed spines are present posteriorly.
The asb spines of the third instar of P. longicornis are dark on the thoracic segments and a1-2, light brown on a3 and colourless on a4-a5. In P. indecora, spines of the asb are dark on thoracic segments and a1-3, and light brown on a4-a5. In the second and third instars of P. heterochaeta, the whole body surface is covered with dark spines or wart-like protuberances. Both spines and wart-like prominences are distinctly coloured and well visible on the body surface as well as on a6-7.
In the first instar, the middle part of segments a1-7 features a transverse crevice (cr) present ventrally (Figure 5f). Elliptical lateral creeping welts (lcw) are covered by minute spines. A bubble The surface of the anal division (ad) is covered with brown pointed spines or blunted wart-like spines (Figures 2c; 5g,h). The anal division ventrally has an anal plate (ap) and subanal papillae (sa) distinctly protuberant in a tube-like form (Figures 5g; 6g). The ap is relatively small and triangular (Figure 6g). A postanal papilla (pa) is indistinct or at most, forming a group of spines on the same level as the adjacent cuticle (Figure 6f,g). The subanal papilla is large, bulgelike and closely appose ap (Figure 5g). Each sa is devoid of spines, equipped with a sensillum basiconicum and two sensilla resembling sensilla ampullacea. In P. longicornis, para-anal papillae (paa) are indistinguishable from the adjacent cuticle (Figure 5g). In P. indecora, the paa are in the form of a fold, laterally to the sa (Figure 6g), while in P. heterochaeta they form distinct cones.
The spiracular field carries posterior spiracles and is surrounded by seven pairs of sensilla (Figures 2c; 5h; 6g). The first instar has papillae p1, p3, p5 in the form of long filamentous processes (Figures 2c; 5h). In the second and third instars, these processes are present in the form of short, yet distinct cones (Figure 6f,g). The remaining papillae (p2, p4, p6 and p7) are indistinguishable from the adjacent body surface on all instars.
Posterior spiracles (ps) are slightly raised above the surface of the ad in the first and second instars (Figure 5h). In the third instar ps are

DISCUSSION
Despite the impact on nestling health and survival, the morphology of Passeromyia larvae has not been comprehensively studied. Limited to two early works (Skidmore, 1985;Zumpt, 1965), previous descriptions are incomplete, which is the case particularly for the finer sclerites of the cephaloskeleton we present here, such as the rami and epistomal and labial sclerites. Skidmore's (1985) examination of the third instar cephaloskeleton attached to the puparium of P. indecora failed to recognize an accessory stomal sclerite and oral bar, most likely due to the preparation method. Previous interpretations of the anal region (Skidmore, 1985) also differ from our observations. The application of SEM enabled the identification of a sensillum basiconicum on the surface of each bulge (sa) lateral to the anal opening. Sensory sensilla are characteristic of subanal papillae and have not been observed in any other anal papillae (Grzywacz et al., 2015). Thus, we conclude that the anal region of Passeromyia larvae protrudes into a tube-like structure and carries apically both an anal plate and subanal papillae.
Larval morphology of Muscidae, particularly details of the cephaloskeleton, provides valuable phylogenetic information and is also a rich source of information on larval feeding habits (Grzywacz et al., 2021;Skidmore, 1985). Our results confirm the position of the genus Passeromyia within Reinwardtiinae on the basis of features shared with other members of the subfamily. These include massive, strongly sclerotized cephaloskeleton, well separated and symmetrical mouth-hooks, broad intermediate sclerite with rod-like paired rami, robust basal sclerite and the distribution and size of accessory oral sclerites and dental sclerites (Grzywacz et al., 2015;Skidmore, 1985;Velásquez et al., 2013). Other features of the larval morphology corroborate the monophyly of Passeromyia, and the most conspicuous of the first instar, which is a unique character state among muscid flies and even within the entire cyclorrhaphan Diptera (Ferrar, 1987). Previous authors did not recognize these processes and their concomitant sensory papillae as homologous with p1, p3 and p5 of other calyptrate flies (Grzywacz et al., 2015). The massive mouth-hooks of first instar Passeromyia is another unique character state within Muscidae (Grzywacz & Pape, 2014;Keilin & Tate, 1930;Schumann, 1954;Skidmore, 1985;Velásquez et al., 2013). The vestigial first instar cephaloskeletal labrum is diagnostic for species of Passeromyia, but is not unique among calyptrate flies. A similar reduction of the first instar labrum is observed in all species of the megadiverse flesh fly subfamily Sarcophaginae, in all species of the Rhinophorinae (now to be considered a blow fly subfamily, see Yan et al., 2021) and in calliphorine blowflies of the genera Bellardia Robineau-Desvoidy, 1863 and Onesia Robineau-Desvoidy, 1830 (Ferrar, 1987). It is therefore difficult to claim a correlation between this character state and a specific feeding habit given the diverse life histories exhibited by the above taxa (Cerretti et al., 2020;Ferrar, 1987;Pape, 1996). However, we propose that the arrangement and colouration of spines on the body segments form species-specific patterns, allowing identification of P. indecora, P. heterochaeta, P. longicornis and possibly P.
steini, pending examination of additional material. Spinal arrangement has been utilized in taxonomic studies of Muscina Robineau-Desvoidy, 1830 (Grzywacz et al., 2015), although this character is not always easily observed.
According to the literature, larvae of the different species of Passeromyia differ in food acquisition strategy (saprophagy, haematophagy, necrophagy), as well as preferred food type (excreta, food particles, blood, dead nestlings) (Skidmore, 1985). Despite the frequently observed relationship between cephaloskeletal structure and feeding habit among dipteran larvae (Ferrar, 1987;Skidmore, 1985), we found that the cephaloskeletons of species of Passeromyia are largely indiscernible, even comparing the saprophagous P. steini with the hematophagous species. For example, unlike the mouth-hooks of other saprophagous Muscidae, the second instar of P. steini is equipped with teeth similar to those used for piercing host skin by obligate parasitic members of the genus (Skidmore, 1985), and this feature is therefore best interpreted as part of the ground plan for the genus. Two alternative hypotheses may explain the lack of morphological diversity among Passeromyia species that occupy different ecological and feeding niches: (1) reported differences in feeding strategy are real and are manifested in either physiological or behavioural adaptations, while cephaloskeleton structures did not differentiate during speciation; (2) reported differences are premature conclusions resulting from a scarcity of field observations. We find the second hypothesis most likely and suggest that larvae of P. steini may be facultative or opportunistic parasites of nestling birds. Observations in the field are required to substantiate this assertion.  (Ferrar, 1979;Grzywacz et al., 2017;Roback, 1951;Skidmore, 1985). Skidmore (1985)   However, an extensive literature search showed that similar rods, named "rami", have been reported in the second and third instars of some Lauxaniidae (Semelbauer & Kozánek, 2011, 2012. To provide greater certainty on these structures, we analysed additional material including some species from the muscid genera Alluaudinella  This study provides the first comprehensive documentation of the cephaloskeleton of muscid species obtained by confocal laser scanning microscopy (CLSM). Despite earlier evidence of the utility of CLSM in visualizing morphological structures, this powerful tool has rarely been used on immature stages of Diptera Li et al., 2021;Szpila et al., 2016Szpila et al., , 2021. The main obstacles are high costs, equipment availability and lack of standard protocols for the preparation and visualization of specimens. The available protocols for the preparation of material for CLSM turn out to be only basic guidelines, and modifications are often required for individual preparations Szpila et al., 2021). Throughout this study, we found that the condition of material, the presence of impurities, previous storage conditions and the time of specimen maceration are critical to a successful CLSM analysis. Nonetheless, CLSM is an innovative tool that allows visualization of the position and shape of fine, complex morphological structures without the requirement for additional staining of specimens .
Here, the application of light microscopy provided insufficient resolution to visualize fine, tightly arranged and multi-layered cephaloskeletal structures of the larvae. The correct interpretation of rami and parastomal bars was possible only thanks to the application of CLSM.
Most importantly, the ability to generate 3D visualizations revealed interactions between individual sclerites that could not be obtained with light microscopy alone.