The early stages and breeding sites of four rare saproxylic hoverflies (Diptera: Syrphidae) from Spain

The puparial stage is described for each of four saproxylic hoverflies (Diptera, Syrphidae): Ferdinandea fumipennis Kassebeer, 1999, Mallota dusmeti Andréu, 1926, Myolepta difformis Strobl, 1909, and Myolepta obscura Becher, 1882. Early stages were collected at Cabañeros National Park in Ciudad Real province, central Spain. Mallota dusmeti, M. difformis, and M. obscura were found in water‐filled holes of live Fraxinus angustifolia Vahl. trees and the larva of F. fumipennis was found in a sap run on a live Quercus pyrenaica Wild. tree. Larvae were identified by rearing them to the adult stage. Puparia of these species possessed generic‐level characters found in congenerics. Apparent species‐level characters separating them from early stages of congenerics are proposed. The records reported here are the first for the Iberian Peninsula of M. obscura. Mallota dusmeti, M. difformis, and Myolepta obscura are extremely rare in Europe with M. dusmeti classed as “vulnerable” in the Spanish Invertebrate Red Data Book.


Introduction
Organisms dependent on dead wood or decaying material associated with the woody parts of trees are known as saproxylic (Speight 1989;Rotheray et al. 2001). Saproxylic Syrphidae are important for biodiversity and conservation in various ways. Some have been proposed as indicators of woodland quality (Morris 1998;Speight 1989). This is because they are restricted to ancient woodland where woodland cover has been continuous for hundreds, possibly thousands, of years and where all growth stages of trees, particularly overmature trees and fallen wood, accumulate under largely natural processes and are mostly left undisturbed or managed using traditional techniques (Reemer 2005). Furthermore, many saproxylic Syrphidae are large and conspicuous and utilize a wide range of dead wood microhabitats. As such they can act as ''flagships'' for the wider community of saproxylic organisms, so that by conserving these syrphids, many additional species also benefit.
Finally, saproxylic syrphids are included in the red lists of many European countries and deserve individual conservation because they are endangered or possess unique ecological and other features over and above their value for site quality assessment and as flagship species. Despite the importance of saproxylic Syrphidae for conservation, protocols for managing ancient woodland to monitor and encourage dependent organisms like saproxylic Syrphidae are little developed (Rotheray and MacGowan 2000). The only saproxylic Syrphidae known to us which are being targeted for individual conservation are Callicera spinolae Rondani, Blera fallax (L.), and Hammerschmidtia ferruginea (Fallén) which are included in the Biodiversity Action Plan Process of the UK.
Utilizing saproxylic Syrphidae for conservation depends on how well they are known. In particular, basic species taxonomy must be worked out and assessments made of each species distribution and status across Europe, within individual countries and within regions of individual countries. Also required are details of adult and larval requirements of which a minimum necessity is knowing the breeding site. Despite the fact that Syrphidae are a popular group of insects and knowledge is increasing across Europe, lack of these basic data is a continuing barrier to progress.
In this paper, we contribute to the ongoing process of providing basic data on saproxylic Syrphidae by describing the puparial stage of four rare species: Ferdinandea fumipennis Kassebeer, 1999, Mallota dusmeti Andréu, 1926, Myolepta difformis Strobl, 1909, and Myolepta obscura Becher, 1882. In addition, breeding sites and the distribution of each species is assessed.

Material and methods
Larvae of M. dusmeti, M. difformis, and M. obscura were found in water-filled holes on overmature, live Fraxinus angustifolia Vahl. tree trunks and larvae of F. fumipennis were found in sap exudations on live Quercus pyrenaica Willd. tree trunks at Cabañ eros National Park (39u239470N, 4u299140W) in the north-west of Ciudad Real province in central Spain.
The park consists of about 40,000 ha of well-preserved Mediterranean ecosystems including a range of woodland types. In the past, many of the woodlands have been managed using traditional techniques including coppicing and pollarding.
Due to lack of taxonomic data, larvae could be identified to genus level, not species level, and were reared in the laboratory to obtain the adult stage for identifying the species. Rearing took place in the dark in a controlled chamber at 20uC, 65-85% humidity. Larvae were reared in plastic boxes with a hole cut in the lid for netting which allowed air exchange, and containing material from the tree hole or sap run in which they were found. Boxes were checked daily to ensure they were sufficiently wet and to remove puparia, which were reared individually on damp tissue in 55 mm diameter Petri dishes and checked daily until adults emerged.
Adults were identified using the keys and descriptions in Andréu (1926), Kassebeer (1999), and Reemer et al. (2004) and checked against named specimens in the collections of the University of Alicante. Descriptions are based on the puparial stage which retains all the features of the third stage larva with the addition of the pupal spiracles (Rotheray 1993). Illustrations were made using a binocular microscope fitted with a drawing tube. Measurements were made using an eyepiece micrometer. Puparium length was measured on the ventral surface from the anterior margin to the anus. Width was the maximum of all structures measured. The length of the posterior respiratory process (prp), or posterior breathing tube, is a useful diagnostic character but it is not measurable along its entire length because in most specimens the base is hidden within the folds of the anal segment.
To overcome this difficulty, comparisons between species were made by measuring the length of the prp above the transverse band to the apex and expressing the result as a proportion of the width of the prp at the transverse band. The transverse band encircles the prp below the apex. It is a characteristic feature of the prp in Eristalinae (Syrphidae) and has various states from a raised ridge, a narrow band of striations or other distinctive surface sculpture to an indented groove (Rotheray and Gilbert 1999). Photomicrographs were obtained using a Hitachi S3000N scanning electron microscope (SEM) operated at 20 kV.
Myolepta puparia were often coated in a white substance excreted by the larva at the time of pupariation. This coating obscured integumental detail and was removed by boiling affected puparia in a solution of KOH for 5 min and picking off loosened deposit with pins. Head skeletons were examined by softening a puparium in a cold solution of KOH for about 30 min and removing them from the antero-ventral margin of the puparium using pins.
Terminology follows Hartley (1961) and Rotheray (1993) for gross morphology and Hartley (1963), Roberts (1970), and Rotheray and Gilbert (1999) for head skeleton structure. Voucher specimens of adult and immature stages of all species have been deposited in the Colecció n Entomoló gica de la Universidad de Alicante (CEUA).

Puparium
Overall appearance. Short-tailed larva with mouthparts of the saprophagous type (Rotheray and Gilbert 1999); truncate anteriorly, slightly tapered posteriorly; prolegs absent, vestiture of fine setae becoming longer towards the anal segment; head skeleton with mandibles and tentorial bar heavily sclerotized ( Figure 2); prp short, slightly longer than wide from transverse band to apex and slightly tapering with a transverse indentation just below the apex ( Figure 2).
Head. Mouth-hooks present but inconspicuous; mandibular lobes partially external (Rotheray and Gilbert 1999) and dorsal lip coated in setae. Head skeleton (Figure 2). Basal sclerite lightly sclerotized; ventral cornu about three times as long as dorsal cornu, bar-shaped with ventral pharyngeal ridges and grinding mill (Hartley 1963) present; dorsal cornu rounded apically; dorsal bridge not heavily sclerotized; in lateral view, tentorial bar about three times wider in front of the tentorial bridge than behind; mandibular sclerite T-shaped with apodeme about twice as narrow as the longer and heavily sclerotized mouthhook.
Anterior spiracles. Anterior spiracles 0.05 mm long by 0.04 mm wide, inconspicuous, lightbrown with four or five oval openings across the apex.
Abdomen. Vestiture of a uniform coating of fine, pale brown setae varying in length from 0.04 to 0.1 mm and becoming longer towards the anal segment; ventral surface of the anal segment with one transverse fold between the anus and the tip of the segment; abdominal segment 7 with sensilla 1-6 on fleshy, rounded papillae; anal segment truncate and inclined dorso-ventrally; first two pairs of lappets on the anterior section of the anal segment and the third pair of lappets on the posterior section; prolegs absent.
Pupal spiracles. Separated by a distance of about 2.5 times their length; each spiracle 0.42 mm long and 0.18 mm wide, brown, sub-cylindrical, inclined postero-laterally; irregularly spaced tubercles bearing spiracular openings extending halfway down the antero-dorsal surface; between tubercles surface smooth but sculptured at base.
Posterior respiratory process ( Figures 5, 6). Short, with visible section about as long as wide; dark brown; matt with a distinct transverse band and a striated indentation below the apex; length between transverse band and spiracular plate 0.35 mm; tapered towards the apex with width at transverse band 0.5 mm; spiracular plates separated by a deep, dorso-ventral indentation; each plate bearing three horse shoe-shaped spiracular openings.
Head skeleton (Figure 8). Basal sclerite at junction of dorsal and ventral cornea black and heavily but unevenly sclerotized; dorsal cornu straight dorsally and tapered posteriorly; ventral cornu inflated posteriorly and 2.5 times as long as dorsal cornu; parastomal bars bordering the epipharyngeal plate clearly visible at the antero-dorsal part of the head skeleton, otherwise sclerotization lacking between the dorsal bridge and the tentorial bar; dorsal bridge and ventral pharyngeal ridges present; ventral bridge linking the tentorial bars broader than tentorial bar; mandibular sclerite T-shaped.
Prothorax. Anterior fold evenly coated in red-brown spicules, translucent at apex. Anterior spiracles (Figure 11). Length50.56 mm; width50.2 mm, yellow brown, less than three times as long as base is wide with pointed and slightly recurved tip; at least 14 spiracular openings present at the apex of the spiracular plate; openings arranged in a curved row.
Mesothorax. Prolegs present with three or more rows of crochets; five crochets in the primary rows.
Abdomen. Vestiture sparse and vestiture absent on ventral surface; prolegs with three rows of crochets present on the first six segments; five crochets in the primary row decreasing to three on segment 6; lateral margin with two aligned projections per segment; segments 1-6 with projections bearing sensilla 4+5 and 6 and segment 7 projections incorporating sensilla 7+8 and 9; anal segment elongate, more than body length, bearing three pairs of approximately equidistant lappets.
Posterior respiratory process. Elongate, shiny brown with each spiracular plate bearing three pairs of curved spiracular openings; transverse band not distinct.

Material examined
Two puparia from larvae collected 19 March 2004 and 28 May 2004, from tree holes containing standing water in live Fraxinus angustifolia and F. angustifolia or Quercus faginea

Puparium
Overall appearance. Short-tailed larva with mouthparts of the saprophagous type (Rotheray and Gilbert 1999); sub-cylindrical in cross-section, truncate anteriorly, tapering posteriorly; prolegs bearing crochets present on mesothorax and first six abdominal segments; anal segment coated in papillae bearing tufts of setae; pupal spiracles club-tipped with basal sculpture of dome-shaped projections; prp narrow and yellow-brown.
Head skeleton (Figure 9). Basal sclerite at junction of dorsal and ventral cornea only heavily sclerotized posteriorly; dorsal cornu rounded dorsally and posteriorly; ventral cornu bar-shaped, not inflated and about four times as long as dorsal cornu; parastomal bars bordering epipharyngeal plate not distinct; sclerotization lacking between dorsal bridge and tentorial bar; dorsal bridge and ventral pharyngeal ridges present; ventral bridge as wide as tentorial bar; mandibular sclerite T-shaped.
Mesothorax. Prolegs present with two rows of crochets.
Abdomen. Vestiture varying in size from 0.03 to 0.12 mm, fine, pale brown setae, except on the ventral surface, where pubescence is inconspicuous; anal segment coated in papillae with tufts of setae at their tips; first and second pairs of lappets indistinct but third pair of lappets conspicuous, 0.28 mm long; prolegs with two rows of crochets present on the first six abdominal segments. Pupal spiracles (Figure 15). Separated by a distance twice their length, 0.46 mm long, yellow-brown in colour, club-tipped with irregularly spaced, rounded tubercles bearing three or four radially arranged spiracular openings on the enlarged apex; base sculptured with dome-shaped projections.
Posterior respiratory process (Figure 16). Length 1.96 mm; width at transverse band 0.22 mm (n53); smooth, without sculpturing, yellowish brown in colour, with an indistinct transverse band and each spiracular plate bearing three pairs of gently curved, spiracular openings, interspiracular setae consisting of three or four long and conspicuous setal branches and a pit between the dorsal and dorso-lateral interspiracular setae.

Puparium
Overall appearance. Short-tailed larva with mouthparts of the saprophagous type (Rotheray and Gilbert 1999); sub-cylindrical in cross-section, truncate anteriorly, tapering posteriorly; prolegs bearing crochets present on mesothorax and first six abdominal segments; anal segment coated in papillae bearing tufts of setae; pupal spiracles tapered, not club-tipped and curved backwards; prp broad and black-brown.
Head skeleton. Similar in detail to M. difformis head skeleton (Figure 9).
Mesothorax. Prolegs present with two rows of crochets.
Abdomen. Vestiture varying in size from 0.04 to 0.09 mm, fine, pale brown setae except on the ventral surface, where pubescence is inconspicuous; anal segment coated in papillae with tufts of setae at their tips; first and second pairs of lappets indistinct but third pair of lappets conspicuous, 0.32 mm long; prolegs with two rows of crochets present on the first six abdominal segments.
Pupal spiracles (see M. difformis spiracles in Figure 15). Separated by a distance twice their length, 0.3 mm long, dark brown except yellow at apex, tapered and curved backwards with irregularly spaced, protruding, and rounded tubercles bearing three or four spiracular openings crowded at the apex of the dorsal surface; sculpture shiny but vaguely indicated dome-shaped projections at base.
Posterior respiratory process (Figure 17). Length 1.96 mm, range 1.92-2 mm; width at transverse band: 0.3 (n52); smooth, without sculpturing, blackish brown in colour, with an indistinct transverse band and each spiracular plate bearing three pairs of gently curved, spiracular openings, interspiracular setae consisting of three or four long and conspicuous setal branches and a pit between the dorsal and dorso-lateral interspiracular setae.

Material examined
Two puparia from larvae collected 25 February 2006, in a wet hole of an east-orientated, live F. angustifolia tree trunk, Españ a, Ciudad Real, Cabañ eros National Park, Fresneda de Gargantilla, two males emerged 15 March 2006.

Discussion
This study is part of a wider project to investigate the Syrphidae of Cabañ eros National Park in central Spain. As with previous studies (Rotheray et al. 2006) we concentrated on finding early stages as well as sampling adults using Malaise traps and other techniques. The value of including early stages in biodiversity and faunistic studies is revealed by the fact that in 2 years of trapping and searching, no adults of three of the four species considered here were obtained (M. dusmeti, M. difformis, and M. obscura).
Mallota species have long-tailed larvae (.body length) in contrast to Ferdinandea species which have larvae with contracted anal segments (,0.5 body length) and Myolepta species which have short-tailed larvae (from half, to as long as, body length). Mallota larvae can be distinguished from other long-tailed syrphid larvae by the smooth integument, lacking or almost lacking vestiture and by sensilla 7-9 aligned and often on a pair of fleshy projections, along the lateral margins of the 7th abdominal segment (Rotheray 1993). The puparium described as Milesia crabroniformis (Fabricius) by Matile and Leclercq (1992) shares these characters but we have examined this specimen and it is Mallota cimibiciformis (Fallén), as was also pointed out by Maibach and Goeldlin- de-Tiefenau (1989). Ferdinandea larvae can be distinguished from other saprophagous syrphids with contracted anal segments by lack of protruding mouthhooks and partially external, fleshy mandibular lobes on either side of the mouth (Rotheray and Gilbert 1999). Myolepta larvae can be distinguished from other short-tailed saprophagous larvae by fleshy papillae bearing tufts of setae that coat the anal segment and setae not spicules on the prothoracic anterior fold (Rotheray 1993). The puparia of the four species described here possess these distinguishing characters which adds support to their value in separating early stages at the generic level.
The early stages of M. dusmeti can be separated from those of M. cimbiciformis-the only other west European species with described early stages (Hartley 1961;Maibach and Goeldlin de Tiefenau 1989)-by the number of openings on the anterior spiracles. There are 14 openings in M. dusmeti but 25 in M. cimbiciformis. Also the lateral ridge bears dark, fleshy projections on all abdominal segments whereas M. cimbiciformis has an inconspicuous ridge and projections are absent except on the 7th abdominal and anal segments. The early stages of an east European species, Mallota eurasiatica Stackelberg, have been described by Sivova et al. (1999). However, it is not possible from the description to separate the early stages of this species. The only other Mallota early stages described are from the Nearctic: M. bautias (Walker) and M. posticata (F.) (Snow 1949in Maier 1978. The puparium of M. bautias apparently differs in having dorsally curved pupal spiracles (Maier 1978). The puparium of M. posticata differs from the other three species by not having a depressed and tapering prothorax and a ''boat-shaped'' profile ( Figure 7). All Mallota species reared to date have been found in tree holes containing standing water. However, in addition to this breeding site, Sivova et al. (1999) record M. eurasiatica from sap under bark of Ulmus trees.
F. fumipennis can be separated from F. cuprea-the only other Ferdinandea species with early stages described (Lundbeck 1916;Hartley 1961;Dušek and Láska 1988;Rotheray and Gilbert 1999)-by the form of the prp and sclerotization of head skeleton. In F. fumipennis, there is a striated transverse indentation between the transverse band and the apex of the prp, whereas in F. cuprea this indentation is absent and the prp above the transverse band smoothly tapers to the apex (Figures 3, 4). The head skeletons of both species are similar but there are differences in the degree of sclerotization of the mandibular sclerite and the width of the tentorial bar anterior to the tentorial bridge. The mandibular sclerite is less heavily sclerotized and the tentorial bar anterior to the tentorial bridge is narrower in F. cuprea (Figures 1, 2).
In addition to F. fumipennis only two other Ferdinandea species have been reared. F. cuprea larvae are found in exudations of sap on the trunks of a wide range of deciduous trees (Rotheray 1993) and in roots of artichokes previously attacked by larvae of Cheilosia vulpina (Meigen) (Brunel and Cadou 1994). F. ruficornis has been reared from a Cossus sap run on Populus (Lundbeck 1916). The breeding record of F. fumipennis supports the idea that Ferdinandea larvae are associated with exudations of tree sap (Rotheray and Gilbert 1999). However, the extent to which other media, such as that found by Brunel and Cadou (1994), are used has yet to be evaluated.
The early stages of the two Myolepta species described here can be distinguished from each other by the pupal spiracles which are club-tipped in M. difformis and straight in M. obscura and from the colour of the prp which is yellowish brown in M. difformis and blackish brown in M. obscura. The only other Myolepta species with early stages described are M. luteola (Gmelin) (Becher 1882;Dušek and Láska 1960;Hartley 1961), M. potens (Harris) (Rotheray 1991) and M. vara (Panzer) (Sivova et al. 1999, figured by Dussaix 1997. The early stages of M. vara apparently differ from those of other Myolepta species by having three not two rows of crochets on the abdominal segments. M. potens differs from other species in having a lateral fringe of setae as long as the lappets on the anal segment and M. luteola differs by having the dorsal pair of spiracular openings at the apex of the prp longer than the other two pairs. In other Myolepta species a setal fringe on the anal segment is absent and the spiracular openings of the prp are more or less equal in length. All reared species of Myolepta have been from tree holes, often small and containing wet decaying sapwood and vegetation: M. obscura in Populus (Becher 1882), Acer, Fagus, and Populus (Dušek and Láska 1960); M. luteola in Fagus and Populus (Hartley 1963); M. potens in Fagus (Rotheray 1991); M. vara in Quercus (Dussaix 1997) and Ulmus (Sivova et al. 1999). The breeding records of M. difformis and M. obscura support previous results that tree holes are the breeding site of Myolepta species.
Syrphids are rapidly becoming established as one of only a few key insect groups used for environmental assessment. Such assessments rely almost entirely on sampling the adult stage, for example using Malaise, sticky, or pan traps (Reemer 2005;Schweiger et al. 2005Schweiger et al. , 2007Dziock 2006;Ouin et al. 2006;Burgio and Sommaggio 2007). These techniques tend to be biased, however, in that they will miss adults of species with restricted flight preferences. Unfortunately, adults of syrphid species that have greatest value for environmental assessment because they are rare and their habitats have suffered from losses and fragmentation, such as saproxylics, appear to fall into this category and are missed. Our results show such a sampling bias. We also show that such biases can be overcome by including searches for early stages. Furthermore, finding and assessing the quality and quantity of breeding sites is a critical requirement required to manage habitats for Syrphidae. To bring out and use syrphids to the full potential as a key assessment group, it is essential that early stages are included.
F. fumipennis was known only from Algeria, Morocco, and Tunisia (Kassebeer 1999) but Ricarte and Marcos-García (forthcoming) have recently recorded this species for Europe with material from the Iberian Peninsula and Balearic Islands. M. obscura has not been recorded previously from Spain (Marcos-García et al. 2002;Speight 2006) and is rare throughout its range of northern France through central Europe, the Balkans, and the Transcaucasus (Reemer et al. 2004). The records presented here extend the western range of this species into Spain. M. difformis has a range similar to F. fumipennis of Algeria, Morocco, and Tunisia, and Spain in Europe (Reemer et al. 2004). Within Spain M. difformis is known from Madrid, Valencia, Cáceres, Cádiz, and Salamanca (Marcos-García 1986). M. dusmeti is known from Spain and also from Tunisia (Speight 2006). In Spain it is classified as ''vulnerable'' in the Invertebrate Red Data book (Marcos-García 2006), being known only from Burgos, Cáceres, Madrid, and Salamanca (Marcos-García 1986). The shared ranges of North Africa and Europe for F. fumipennis, M. difformis, and M. dusmeti demonstrate a clear faunal link between these continents in terms of Syrphidae. The fact that the most northerly point of distribution of these species is Spain makes their conservation particularly important in that country. For this to be achieved, a key task will be the preservation of overmature trees where larvae of these species develop. Sympathetic management of woodlands with special attention to old trees would allow protection of these Syrphidae and other dependent saproxylic insects that have yet to be assessed in Spain. As our results, and also those of others such as the newly discovered weevil species (Coleoptera, Curculionidae) breeding in saproxylic macrofungi (Micó , personal communication) show, the potential for new discoveries remains high in Spanish woodlands.