© Entomologica Fennica. 5 December 2007

Description of immature stages of cleptoparasitic coecutiens and trochantericus (: : , Protepeolini) with remarks to their unusual biology

Jakub Straka & Petr Bogusch

Straka, J. & Bogusch, P. 2007: Description of immature stages of cleptoparasitic bees Epeoloides coecutiens and Leiopodus trochantericus (Hymenoptera: Apidae: Osirini, Protepeolini) with remarks to their unusual biology. — Entomol. Fennica 18: 242–254. This study presents descriptions of first instars and mature larvae of clepto- parasitic bees Epeoloides coecutiens (host fulvipes)andLeiopodus trochantericus (host Diadasina sp.). Larvae of the Epeoloides are de- scribed for the first time. The of Epeoloides is similar to other osirine spe- cies, but also possesses several characters previously known as autapomorphies of the genus Leiopodus. The unusually short stadium of the first instar and kiling behaviour of the second instar in E. coecutiens are recorded. The first instar of L. trochantericus is quiescent for a long time and kills the late second or third instar of the host. J. Straka, Faculty of Science, Charles University in Prague, Department of Zool- ogy, Praha, Czech Republic; E-mail: [email protected] P. Bogusch, Faculty of Science, Charles University in Prague, Department of Philosophy and History of Science, Praha, Czech Republic; E-mail: boguschak @seznam.cz Received 23 November 2006, accepted 27 April 2007

1. Introduction Mature larvae of at least one of all the totally cleptoparasitic tribes of the family Apidae Nest is a life strategy of cleptoparasitic have already been described, but the number of bees that evolved a number of times de novo genera with undescribed morphology is still large within the superfamily Apoidea. Recent studies (Rozen 2001, Rozen et al. 2006). For this reason, suggest six or eleven independent origins among Rozen’s (2001) taxonomic key to the genera of the bees of the family Apidae (Michener & cleptoparasitic larvae is highly incomplete. Greenberg 1980, Rozen 2000, Michel-Salzat et This key contains all the known genera except the al. 2004, Straka & Bogusch 2007). Both opinions most recently described species of the tribe are supported by phylogenetic analysis based on Osirini (Rozen et al. 2006). Thus, members of adult and larval morphology respectively (Roig- this tribe are completely missing in the key. Alsina & Michener 1993, Straka & Bogusch Generalized strategies of egg laying and host 2007) and by the diversity of the modes of para- offspring killing are known for all of totally sitism (Rozen 1991, Rozen et al. 2006). cleptoparasitic tribes of the subfamily and ENTOMOL. FENNICA Vol. 18 • Immature stages of cleptoparasitic bees 243 almost all tribes of the subfamily and discuss the modes of parasitism of Epeo- (Rozen 1991, Rozen et al. 2006). This knowledge loides coecutiens (Fabricius) and Leiopodus enables a wide comparison to be made of the vari- trochantericus Ducke. Larvae of the genus Epeo- ous lineages of bees and currently described spe- loides have never been described, and parasitic cies of the tribes Osirini and Protepeolini. The behaviour is known for only one species in the most important behavioural questions are: (1) are tribe Osirini (Rozen et al. 2006). In contrast, lar- eggs introduced into open or closed host cells, vae of the tribe Protepeolini are well known and and (2) does the mother kill the host egg or early the description of those of L. trochantericus instar with her mandible or sting, or do the para- makes the list of described species almost com- sitic larvae kill the host egg or larva? (3) If the lat- plete. Five species of the genus Leiopodus are the ter, which parasitic larval instar(s) does the kill- only representatives of the tribe and the larvae of ing? four of them are currently known (Roig-Alsina & Mature larvae of two species of the tribe Rozen 1994). Osirini (Protosiris gigas Melo and Parepeolus minutus Roig-Alsina) were described by Rozen et al. (2006) with detailed information about the 2. Material and methods mode of parasitism and first instar morphology for Protosiris Roig-Alsina. Both osirines are nest Larvae of Epeoloides coecutiens were collected parasites of oil collecting bees of the tribe from 25.VII.2005 to 20.VIII.2005 in Klánovický . Protosiris opens finished closed les, Praha–Klánovice in the Czech Republic, by host cells and kills the host egg or larva with its both authors. The sample contains three dead first elongate sting. Similar behaviour is known only instar larvae, two killed by intraspecific siblings, in the unrelated Hoffmannsegg (Garó- and one died shortly after eclosion in laboratory; falo & Rozen 2001). Larval instars of Epeoloides two second instar larvae; three undefecated Giraud were not previously described, but there fourth instar larvae; and three mature (post- are some publications directed towards the be- defecating) larvae. haviour of adults (Bogusch 2003, 2005). Bo- Larvae of Leiopodus trochantericus were col- gusch (2005) mentioned an important feature of lected from 29.I.2006 to 31.I.2006 on the road cocoon spinning in Epeoloides, which indicates close to Capitan Solari, Chaco province, in Ar- development of a pronounced labiomaxillary re- gentina by the first author. One first instar larva gion in its mature larvae. and three mature larvae represent the sample. Mature larvae, first instar larvae and modes of Other biological observations were performed parasitism of two species of Protepeolini [Leio- also on the nesting area on the road close to podus singularis (Linsley et Michener) and L. Mburucuyá, Corrientes province, in Argentina. lacertinus Smith] have already been described by All specimens are deposited in the collection of Rozen et al. (1978) and Roig-Alsina and Rozen the first author. (1994). Roig-Alsina and Rozen (1994) have also Biological observations were made on the described predefecating larva of L. abnormis same localities where the larvae ware collected. (Jörgensen), but it seems to be very similar to L. Construction of the cell and the shape of the pro- singularis. Bees of the tribe are the vision of Macropis fulvipes (Fabricius) were very only known hosts of Leiopodus Smith. Females similar to M. nuda (Provancher) (see Rozen & Ja- enter their opened cell while it is still being cobson 1980). The food mass was formed into provisioned and precisely embed their egg in the compact ovoid loaf and was positioned toward cell wall. In L. singularis the first instar hatches a the back of the cell. Egg was placed on the top of long time after the host larva and usually kills it the mass. Laboratory rearing of larvae was when it is quiescent before molting to the third or performed with plastic containers of various fourth instar (Rozen et al. 1978, Roig-Alsina & shapes in the laboratory at room temperature (20– Rozen 1994). 25°C). The goal of the current study is to describe All specimens were preserved directly in mature and first instar larvae as well as to describe Pampel’sfluid(glacialaceticacid–4parts,5% 244 Straka & Bogusch • ENTOMOL. FENNICA Vol. 18 formaldehyde – 6 parts, 96% ethyl alcohol – 15 lateral view (Fig. 1c). This character appears like parts, distilled water – 30 parts) according to the feature in Protepeolini, but there the extension Švácha and Danilevsky (1987). of hypostomal ridge is more internalized (see the Mature larvae, first instar larvae (one of each Discussion). Also the conspicuous body spicu- species) and second instar larva (Epeoloides lation and distinct setae on the head capsule de- only) were dissected and cleared in a 10% solu- note specific features (autapomorphy) of the tribe tion of potassium hydroxide (KOH) similarly to Protepeolini. However, dorsal body spicules are Rozen (2001). Head capsules were severed from simple in Epeoloides, not transverse and serrate the rest of the body and cleared in KOH until the like in Leiopodus. soft tissue was removed. Right mandibles were In contrast to Protepeolini, Epeoloides larva removed from the cleared head capsules and ob- possesses some other characters of Osirini; short, served separately from dorsal and inner or outer non-membranous clypeus, upper margin of views. Several spiracles were removed from the labrum well below the level of anterior tentorial right side of the body wall and cleared in the same pits (in frontal view), lower margin without way as the head capsules. Head features, spiracles paired tubercles and large spiculate hypo- and all integumental characters were investigated pharynx. Several mandibular and labiomaxillary in glycerine preparations, sometimes with cotton characters of Epeoloides are similar to those of wool for better adjusting of the right position. Rhathymus Lepeletier et Serville, Cleared specimens were stored in glycerine. Ma- Latreille, and some other , but differ by ture larvae were described according to the characters mentioned above. McGinley (1981) and some more recent works Description. Length (if straight) approxi- (Michelette et al. 2000, Alves-dos-Santos et al. mately 12–13 mm (N = 3). 2002). Descriptions of the first instar larvae are Head: Integument unpigmented except the patterned after those in a major work about first following areas (see Fig. 1c): mandibular apex instar larvae (Rozen 1991). and articulations, ventral extension of hypo- Drawings were prepared using a drawing tube stomal ridge and tentorium in the vicinity of ante- (camera lucida) attached to an Olympus BX40 rior and posterior tentorial pits, apex of labrum, light microscope (from temporary glycerine salivary lips, labial and maxillary palpi, faintly slides) and an Olympus SZX9 binocular micro- pigmented antenna, cardo and stipes; cranium, scope. Any asymmetry shown in the drawing but clypeus, labrum, apices of labium and maxilla not mentioned in the respective description is with scattered setiform sensillum; apex of caused by artefacts of preparation. Drawings labrum, hypopharynx and maxilla with well de- were scanned using an hp scanjet 5530 veloped spicules, epipharynx without spicu- photosmart scanner and refined using Adobe lation. Photoshop 7.0. Final figures were treated using Head size small compared with body (Fig. CorelDRAW 9. 1a); head capsule wider than long in frontal view (Fig. 1b). Tentorium complete; dorsal arm thin and hardly visible; posterior arm well developed, 3. Results projecting medially, lower in position; anterior tentorial pit high, well separated from mandible. 3.1. Postdefecating larva Median longitudinal thickening of head capsule of Epeoloides coecutiens (Fig. 1) well developed, reaching up to middle of head capsula. Postoccipital ridge very slightly curving Diagnosis. Epeoloides is similar to both de- forward (almost straight), well defined in all parts scribed Osirini and in some ways resembles by pigmented internal ridge; hypostomal ridge members of Protepeolini, Melectini, and Rhathy- well developed, sinuate, binding to postoccipital mini. It differs from all other species due to a ridge at obtuse angle; ventral extension of hypo- unique combination of characters and especially stomal ridge externally visible in lateral view due to the morphology of the hypostomal ridge; (Fig. 1c), in some specimens is not pigmented, its posteroventral extension is well visible from a but is recognisable as ventrally salient thickening ENTOMOL. FENNICA Vol. 18 • Immature stages of cleptoparasitic bees 245

Fig. 1. Epeoloides coecutiens, mature larva. – a. Entire larva in lateral view. – b. Head in frontal view. – c. Head in lateral view. – d. Right mandible in dorsal view. – e. Right mandible in inner view. –f.Spiracleinside view. of posterior part of hypostomal ridge (only in maxillary region (Figs 1b and 1c) moderately re- cleared head capsules); junction of hypostomal cessed, enlarged only during cocoon spinning. ridge distinct and pigmented; pleurostomal ridge Maxillary apex well separated from labium, bear- faintly developed; epistomal ridge well devel- ing elongate palpus apically; galea not evident; oped, especially its lateral sections. Parietal band cardo and stipes well developed, distinct as pig- represented by weak oblique line above antenna. mented sclerites. Labium divided into pre- Antennal prominence moderately weakly devel- mentum and postmentum (Fig. 1c); premental oped; disk differentiated from papilla; antennal sclerite faintly evident; labial palpus elongated, papilla distinctly projecting, height about one and slightly shorter than maxillary palpus. Salivary half basal diameter, with approximately seven lips projecting, sclerotized; slightly narrower sensilla. Short labrum sloping straight down- than distance between labial palpi. Hypopharynx ward, bearing numerous conspicuous sensillum, (Fig. 1b) large, produced dorsally, apical part labral apex pigmented; clypeus slightly project- slightly blobbed; hypopharyngeal groove well ing beyond labrum and frons (Fig. 1c); developed. epipharynx a simple, slightly curved surface be- Body: Integument without obvious setae al- neath labrum. though those of perianal area with scattered, very Mandible (Figs 1d and 1e) moderately robust, fine setiform sensillum; integument under high short, hardly reaching midline of head; mandibu- magnification distinctly spiculate, patches of lar apex simple, broadly rounded; apical concav- fine, evenly spaced spicules visible on dorsal as ity well developed, smooth, scoop-shaped; upper well as ventral areas of thorax; integument with- and lower apical edges with distinct teeth (Fig. out any spines or sclerotized tubercles. Body for- 1d); outer surface bearing numerous setiform m (Fig. 1a) moderately robust; intersegmental sensilla; cusp slightly distinct, nondentate. Labio- lines moderately defined; dorsal intrasegmental 246 Straka & Bogusch • ENTOMOL. FENNICA Vol. 18

Fig. 2. Epeoloides coecutiens, first and second instar. – a. Entire first instar in lateral view. – b. Head, first instar, in frontal view. – c. Head, first instar, in lateral view. – d. Head, first instar, in ventral view. – e. First instar, right mandible in dorsal view. – f. First instar, right mandible in outer view. – g. Second instar, right mandible in dorsal view. – h. Second instar, right mandible in outer view. lines evident on most body segments; small of body setae. It is very similar to the first instar of paired dorsal tubercles well developed on tho- Leiopodus due to the short pad-like maxillary pal- racic segments, becoming less evident towards pus. The Large swollen labrum with two apical abdominal segment VIII; middorsal and lateral labral tubercles is directed posteriorly and placed body tubercles missing; abdominal segment IX between mandibles and the buccal cavity, which with venter not produced, so that segment X ap- seems to be a unique character of Epeoloides. pears to be attached approximately to middle of Description. Length: 1.7–1.8 mm (N = 3). IX in lateral view; anus in position slightly dorsal Head: Shape (Fig. 2c) hypognathous, pari- because venter of abdominal segment X is some- etals globose, extending upwards and backwards; what more produced than dorsum; perianal area in frontal view (Fig. 2b) approximately as wide as with sclerotized transverse line or ill-defined long; foramen widely opened, almost as wide as ridge. Spiracles (Figs 1a and 1f) moderately the head capsule; head not pigmented except the small, subequal in size; atrial wall without ridges; posterior tentorial arm and apex of mandibles. atrium globular, projecting slightly above the Head capsule sensilla short. Anterior tentorial pit body wall; rim well developed; peritreme moder- conspicuous; posterior tentorial pit darkly pig- ately wide; primary tracheal opening with collar; mented; extension of hypostomal ridge extending subatrium short, usually with 8 chambers. posteroventrally. Postoccipital and hypostomal ridges well developed; pleurostomal ridge faintly developed, extending medially as lateral section 3.2. First instar of Epeoloides coecutiens of epistomal ridge; median section of epistomal (Fig. 2a–f) ridge faintly developed; pale median ecdysial line weakly developed; posterior margin of head cap- Diagnosis. First instar of Epeoloides is simple to sule forms obtuse angle with hypostomal ridge (if recognise due to the combination of following tentorial pit is excluded). Parietal band not evi- characters: hypognathous head (Fig. 2c), unmod- dent. Antennal prominence weakly developed, ified abdominal segment X (Fig. 2a) and absence fused with parietal; disk not differentiated from ENTOMOL. FENNICA Vol. 18 • Immature stages of cleptoparasitic bees 247 papilla; antennal papilla slightly projecting, with paea Warncke was also present in the locality numerous non-setiform sensilla. Labrum broad (but less common). and perhaps large (clypeal margin not completely Behaviour of adults and egg deposition: obvious), strongly curving towards buccal cavity, Flight activity occurs from approximately 9:00 with transverse irregular rows of tubercles on a.m. to 7:00 p.m. (Central European Time). Fe- each side (projections of sensillary basement) males searched for the host nests only in warm which give illusion of labral apex (Figs 2b and weather (23–37°C) and usually only in the sunny 2c); these sensillary tubercles are asymmetrically parts of the nesting area. When the temperature distributed, each bearing one very small apical fell (20–25°C) the bees flew strictly in shafts of papillate sensillum; possible true apical tubercles sunlight or they were inactive. Only during the (Fig. 2d) widely separated from each other, pro- warmest days were they able to fly in the shade jecting posteriorly between mandibles and buccal for a longer time. The parasitic female was usu- cavity. ally able to find all active host nests at the aggre- Mandible (Figs 2e and 2f) broad at base, api- gation, even if the host nests were several meters cally attenuate, sharply pointed, curved into the apart. Females returned to the nest up to ten times buccal cavity when closed, inner edge smooth, during the day and checked the cell condition with ten setiform sensilla on outer surface. Labio- (opened, provisioned, or closed). Some females maxillary region (Fig. 2d) unsclerotized; maxilla entered the nest to check the cell condition imme- well separated from labium, bearing numerous diately after finding it. small spicules; maxillary palpus flat and short, This behaviour often resulted in conflict if the slightly pad-like, born ventrally, bearing several host female was inside. Parasitic females flew small sensilla; maxillary lobe small; labium cov- away immediately after the first contact with the ered with short, posteriorly directed spicules; api- hosts. Areal conflict situation happened when the cally positioned labial palpus distinctly project- host female returned from foraging trip and the ing, conical, short. Hypopharynx (Fig. 2d) large, Epeoloides female had been just inside. The bees densely spiculate; hypopharyngeal groove well fought for several minutes and then the parasite distinct; salivary opening hardly visible, well escaped. In another type of visiting behaviour, the separated from hypopharyngeal groove. parasitic female waits five to twenty centimetres Body: Form (Fig. 2a) moderately elongate, from the entrance on grass or other elevated straight, without distinct tubercles; abdominal points (they usually used our nest markers), until segment X rounded, without apicolateral lobes the host female returns from foraging and subse- (pygopods), but with distinct vertical ridge in quently leaves the nest. As soon as the host fe- dorsal lobe dividing it into two swellings. Integu- male flies away, the bee enters the nest. ment without any conspicuous setae; all body Two females together were often observed in segments with distinct band of spicules directed such behaviour in front of a nest. posteriorly; segment X with rounded spicules If the host nest was inappropriate for egg lay- dorsally and without spicules ventrally; dorsal ing, the female stayed inside several seconds to spiculation sparser and shorter than ventral five minutes. When the nest was appropriate, the spiculation. All spiracles present, normal in posi- cleptoparasitic female stayed one hour or longer tion, subequal in size. Anus positioned apically. inside. Sometimes such a female was seen in the entrance taking the soil from tumulus when fin- ishing the cell closure. Also the closing of the nest 3.3. Biology of Epeoloides coecutiens by two females was twice observed. One was (Figs 2g–h) more inside near the cell and the second closed the middle part of the tunnel with soil. Epeoloides We did all observations in a small area with three female was usually closing the cell as well as the nesting aggregations with one to nine nests of large part of tunnel, when enough soil was avail- host bees (depending on time and activity). able from the tumulus. Macropis fulvipes (Fabricius) was recorded as the Successful parasitation of the nest was artifi- only host of Epeoloides, even though M. euro- cially initiated five times. Elimination of host fe- 248 Straka & Bogusch • ENTOMOL. FENNICA Vol. 18 male during good weather, when this host female was excavated. Three cells were found inside. remained more than one hour at the entrance and One, freshly parasitised, contained two eclosed did not forage, always caused subsequent suc- egg chorions adhered to the bottom right of the cessful parasitation. Freshly closed host cell (M. cell wall, two fresh dead (probably killed) first in- europaea) was not accepted by the female of the stars close to the eggs and two second instar lar- cleptoparasite, but such a situation was observed vae on the top of the provision. One of these sec- only once. In no other case we were certain if the ond instar larvae was decapitated and being eaten cell was really freshly closed and subsequently by the other second instar. These killing larvae visited by the cleptoparasite. had sharp mandibles almost like the first instar 25.VII. – one nest excavated thirty minutes (Figs 2g and 2h). There were no vestiges con- after the entry by Epeoloides female: Two fe- nected with the presence of host larva. males were found inside, one was burrowing and Cocoon fibre of Epeoloides is ochraceous and one was closing the nest. One fully provisioned thus very slightly differentiable from the dark (or nearly so) cell was found with one Epeoloides brown host cocoon. egg. This egg was laid between provision and clo- sure, adhering to the bottom right side of the cell wall. Host egg was not found, probably missing. 3.4. Postdefecating larva 28.VII. – one nest excavated after departure of Leiopodus trochantericus (Fig. 3) of Epeoloides female: One provisioned cell was found with one Epeoloides egg. This egg had al- Diagnosis. Larvae of three other species of the most the same position as in the previous cell. genus Leiopodus have already been studied. Host egg was not found, probably missing. Leiopodus trochantericus differs from them in 18.VIII. – one nest excavated after Epeo- the following characters: maxillary spiculation loides female leavs: One provisioned cell was absent, mandible attenuate apically and its apical found with one Epeoloides egg. Normal position concavity weakly developed, body integument of the egg was disrupted, but it was not on the pro- apparently nonsetose, body spiculation distinctly vision. Host egg was not found, probably miss- shorter than in other species, spiracular atrial wall ing. smooth and subatrial length short. Development cycle and larval behaviour: In Leiopodus trochantericus is a typical member two nests, where the time of egg deposition was of the tribe Protepeolini and in the key to mature known, the time period of larval cycle were re- larvae of cleptoparasitic bees (Rozen 2001) goes corded. Both larvae were dug out together after to the genus Leiopodus (couplet 11). 22 and 27 days respectively and both larvae were Description. Length (if straight) approxi- still feeding. They were reared in the laboratory mately 12–13 mm (N = 3). and the older one spun a cocoon 33 days after Head: Integument unpigmented except for parasitation of the nest. The second one fell from following areas: mandibular apex and articula- the provision and after one day without provision tions, internal head ridges, posterior extension of began spinning of the cocoon as well. Its larval the hypostomal ridge, two vertical lines mesad of development took 28 days. antennae, salivary lips and stipes; labial and The nest with the egg from 28.VII. was kept in maxillary palpi and antennae faintly pigmented; the laboratory until the first instar emerged. The cranium, labrum and apex of labium with scat- first instar hatched overnight in an interval of be- tered setiform sensillum; spiculation not devel- tween 24 and 34 hours when the checks were oped. made. Similar observation was made in the field. Head size small compared to the body (Fig. A nest that we excavated during the evening on 3a); head capsule very wide, distinctly wider than 1st August was provisioned during 28th July and long in frontal view (Fig. 3b). Tentorium com- probably finished (and parasitised) the next day plete; dorsal and anterior arms thin and hardly (29.VII.). One early second instar feeding on top visible; posterior bridge very thin, lateral part of of the provision was found inside. On 5.VIII., ex- bridge is more probably posteroventral and inter- actly 72 hours after the parasitation, a third nest nalized extension of hypostomal ridge (see Dis- ENTOMOL. FENNICA Vol. 18 • Immature stages of cleptoparasitic bees 249

Fig. 3. Leiopodus trochantericus, mature larva. – a. Entire larva in lateral view. – b. Head in frontal view. – c. Head in lateral view. – d. Right mandible in dorsal view. – e. Right mandible in inner view. cussion and Fig. 5b); anterior tentorial pit low, nous area between frons and labrum developed; much closer to mandible than to antenna (Fig. frons with two heavily sclerotized vertical ridges 3b). Median longitudinal thickening of head cap- mesad of antennae of uncertain origin; epi- sule well developed, reaching middle of the head pharynx forms separate large swollen area be- capsule. Postoccipital ridge straight, well defined neath the labrum clearly visible in frontal view in all parts; hypostomal ridge well developed, (Fig. 3b). sinuate, binding to postoccipital ridge at obtuse Mandible (Figs 3d and 3e) robust, very short, angle; long posteroventral extension of hypo- attenuate towards the apex; mandibular apex sim- stomal ridge visible in lateral view including ex- ple, acute; apical concavity poorly developed, ternal and internal parts (see description of smooth, not scoop-shaped; upper and lower api- tentorium and figs 3c and 5b); pleurostomal ridge cal edges with large teeth (Fig. 3d); outer surface well developed; epistomal ridge short, well de- bearing several setiform sensillum; cusp not evi- veloped only laterally. Parietal band not distinct. dent. Labiomaxillary region (Figs 3b and 3c) Antennal prominence moderately weakly devel- large in relation to the head capsule; maxilla mod- oped; disk differentiated from the papilla; an- erately fused to labium at base; maxillary apex tennal papilla distinctly projecting, height about produced medially; palpus large, positioned api- one and half basal diameter, with approximately cally; galea not evident; cardo reduced, stipes dis- six sensilla. Labrum large, rounded, but not pro- tinct as pigmented sclerite. Labium divided into jecting as seen in lateral view (Fig. 3c), bearing prementum and postmentum (Fig. 3c); premental numerous conspicuous sensilla; labral apex with- sclerite faintly evident; labial palpus small, much out tubercles, slightly emarginated ventro- smaller than maxillary palpus, considerably medially; clypeus extremely short laterally, fus- lower than salivary opening. Salivary lips large, ing with frons above, but the frontoclypeal mor- projecting, sclerotized; much narrower than dis- phology is not fully understood, short membra- tance between labial palpi. Hypopharyngeal area 250 Straka & Bogusch • ENTOMOL. FENNICA Vol. 18

Fig. 4. Leiopodus trochantericus, first instar. – a. Entire larva in lateral view. – b. Head in frontal view. – c. Head in lateral view. – d. Head in ventral view. – e. Right mandible in dorsal view. – f. Right mandible in outer view. not differentiated from dorsal surface of labium; 3.5. First instar of Leiopodus trochantericus hypopharyngeal groove absent. (Fig. 4) Body: Integument with obvious pigmented setae on abdominal segment X; lateral part of ma- Diagnosis. The first instar of L. trochantericus is jority of segments with a few very short setiform similar to both described first instars of the genus sensilla visible only under high magnification, Leiopodus. This species looks like the first instar body not conspicuously setose; integument dis- of L. singularis with a more circular head, nar- tinctly spiculate, patches of fine, often transverse rower labrum and flat, padlike maxillary palpus spicules visible on dorsal as well as ventral areas but it differs because of a wider head (but not so of thorax; integument without any spines or wide as in L. lacertinus) and sparsely spiculate sclerotized tubercles. Body form (Fig. 3a) moder- mandibles. Also the body shape, head capsule ately robust; intersegmental lines not defined, not (lateral view), mandible and antennal promi- divided into cephalic and caudal annulets; small nence differ from both described species. paired dorsal tubercles poorly developed on tho- Description. Length: 1.1 mm (N = 1). racic segments; small middorsal body tubercles Head: Shape (Fig. 4c) more or less hypo- distinct between thoracic and anterior abdominal gnathous, parietals globose, extending upwards segments; lateral body tubercles missing; abdom- and backwards; in frontal view (Fig. 4b) wider inal segment IX with non-produced venter, but than long; foramen magnum very large; mandi- segment X attached dorsally to IX in lateral view; bles and head capsule darkly pigmented except anus slightly dorsal in position because venter of the frontoclypeal area and labrum; labiomaxillary abdominal segment X is somewhat more pro- region entirely membranous and unpigmented. duced than dorsum; perianal area with distinct Head capsule sensilla long, setiform. Anterior sclerotized transverse ridge. Spiracles moder- tentorial pit conspicuous; posterior tentorial pit ately small, subequal in size, born on small tuber- internalized. Hypostomal ridge strongly curving cles; atrium globular, projecting distinctly above upward to the postoccipital ridge, both with fluent the body wall; atrial wall without any ridges or conversion from one to another; extension of spines; rim poorly developed; peritreme moder- hypostomal ridge very large, internalized and di- ately narrow; primary tracheal opening with a rected posteroventrally, darkly pigmented; collar; subatrium short. pleurostomal ridge well developed, extending ENTOMOL. FENNICA Vol. 18 • Immature stages of cleptoparasitic bees 251 medially as the lateral section of the epistomal 3.6. Biological observations ridge; median section of the epistomal ridge not of Leiopodus trochantericus developed; pale median ecdysial line present. Pa- rietal band indistinct. Antenna located on large Several females were observed in the nesting area lateral swelling; antennal papilla and disk not dis- of Diadasina (Diadasina) sp. in Chaco prov. as tinct, fused with parietal; swelling bears six or well as in Corrientes prov. Females usually en- seven large nonsetiform sensilla. Labrum large, tered the nest soon after the host left it. They were moderately wide, strongly curving towards often found sitting on the ground near the active buccal cavity, with seven large sensillary tuber- nest and waiting until the host female left. From cles (Fig. 4d); these sensillary tubercles distrib- nine excavated cells, five cells of Diadasina were uted into two lines, each bearing one very small parasitised. A first instar larva was found half in- apical papillate sensillum; apical tubercular pair side a decapitated host second (or possibly third) directed towards the buccal cavity; area between instar. All mature larvae were found in cocoons. these tubercles distinctly emarginated. The cocoon is ochraceous, superficially hardly Mandible (Figs 4e and 4f) broad at base, api- differentiable from the host cocoon. cally attenuate, sharply pointed, curved into buccal cavity when closed, inner edge smooth, with three large and a few small spicules on outer 4. Discussion surface. Labiomaxillary region (Figs 4c and 4d) not sclerotized; maxilla distinct from labium, In the key to the larvae of cleptoparasitic bees bearing long, but very thin setiform posteriorly Epeoloides goes next to Rhathymus (couplet 17) directed spicules; large, flat and padlike (Rozen 2001), but other Osirini were described maxillary palpus located ventrally, bearing six later so they are all missing from this key. Re- distinct sensilla; maxillary lobe large, slightly cently described Osirini Protosiris and Parepe- pronounced over the labium; labium covered olus, (Rozen et al. 2006) go to the couplet 17 as with short, posteriorly directed spicules; labial well as Epeoloides. The partial similarity of palpus may be represented by large short sensilla Epeoloides to Leiopodus is not important for the distributed over lower surface of labium, with keying, because they differ in many other charac- several concentrated on the bare area (Fig. 4d). ters that are more easily visible. Their common Hypopharyngeal area reduced; salivary opening features, like conspicuous setae on head capsule, distinct, directed into buccal cavity. distinct body spiculation, and externally visible Body: Form (Fig. 4a) moderately robust, extension of hypostomal ridge, may be phylogen- slightly curved, with small lateral tubercles below etically important, because all of them were pre- spiracles; abdominal segment X hardly visible, viously considered as unique for Leiopodus probably small, rounded, without distinct apico- (Roig-Alsina & Rozen 1994). On the other hand, lateral lobes (pygopods). Integument with con- body spiculation basically differs between these spicuous setae spread over lateral tubercles; all two genera. Especially the dorsal spicules of body segments with distinct band of spicules di- Leiopodus are distinctly serrate and transverse, rected posteriorly; spicules long on thoracic seg- which is a unique shape of spicules among Long- ments I and II; thoracic segment III and abdomi- Tongued bees ( and Apidae). In ad- nal segments I–VIII with rather short and some- dition, the extension of the hypostomal ridge is times indistinct spicules; spicules of segment IX similar only externally (compare figs 1c and 3c). and X long and distinct, directed posteriorly; dor- The internalized part of the hypostomal ridge of sal spiculation longer than the ventral spiculation. Leiopodus is extremely large and postero- All spiracles present, normal in position, sub- ventrally moved, unlike any other bee species equal in size. Anus not evident. (compare figs 5a and 5b). Roig-Alsina and Rozen (1994) discuss this unusual character and consid- er that the posterior arm of the tentorium is the same as the posterior extension of the hypostomal ridge. We are convinced that this is right and we 252 Straka & Bogusch • ENTOMOL. FENNICA Vol. 18

Fig. 5. Foramen magnum of mature larvae, distal view. – a. Epeoloides coecutiens, position of hidden hypostomal ridge indicated by arrow. – b. Leiopodus trochantericus. use the general term “extension of hypostomal Information about the most important behav- ridge”, which could be internal, external, and var- ioural features of Epeoloides coecutiens has still iously directed (ventrally, posteriorly, etc.). The not been favourably resolved. Several observa- distinctive, darkly pigmented junction of the hy- tions reflect the unusual behavioural strategy of postomal ridge to the tentorium, and in Epeo- this cleptoparasitic bee. Females, when laying loides also to the labiomaxillary region supports eggs, stay inside the appropriate host nest for a this statement. Homology is also indicated by very long time. Construction of the closure is branching of the anterior tentorial arm to the pos- clearly a major part of this time, but the remaining terior tentorial bridge close to the junction (Figs behaviour is unknown. The situation found in ex- 5a and 5b). cavated nests indicates two almost equally prob- The first instar of Epeoloides seems to be able modes of behaviour. First, they are laying more similar to Leiopodus than to Protosiris ac- their eggs in unfinished cells with enough provi- cording to Rozen’s table of derived first instar sion. The second possible strategy is the female characters (Rozen 1991, Rozen et al. 2006). cleptoparasite opening freshly closed cells and Epeoloides shares four possible synapomorphies consuming host eggs. Thus, host eggs were not with Leiopodus: parietals extending upwards and found in any opened host cells. Both possible egg backwards; angle of posterior margin of head to depositions are unique among bees of the family the hypostomal groove is broadly curved; an- Apidae. Consuming of host eggs is known in tenna fused with head capsule; and maxillary pal- cleptoparasitic bees of the families Halictidae and pus short, large and pad-like. Only two derived Megachilidae (Michener 2000) where two or characters are similar to the only known osirine more eggs do not occur in a host cell together. first instar (Protosiris): the egg is free in cell; and Halictid parasitic bees probably destroy the eggs the antenna is fused with head capsule. It is not of other conspecifics as well, which is probably substantial evidence of the relationship between not true in the case of Epeoloides. In one exca- Leiopodus and Epeoloides, but this conclusion is vated nest we found four larvae per cell. In this similar to the result of the cladistic analysis of lar- point of view, the first propounded suggestion is val characters (Straka & Bogusch 2007) Leio- slightly more likely than the second. Observed podus and Epeoloides may be more related than behaviour could be a halfway strategy between Epeoloides to Protosiris or other osirine bees. the opening of finished closed cells and entering The mentioned morphological characters should unfinished host cells without finished pro- not be behaviourally dependent , be- visioning. Until now, these were thought to be the cause the gap between the behaviour of only two possibilities. Leiopodus and Epeoloides is very large (see dis- Other unusual features of Epeoloides are rep- cussion below). resented by the very brief duration of the first ENTOMOL. FENNICA Vol. 18 • Immature stages of cleptoparasitic bees 253 instar followed by the molting to the second and might have had a common cleptoparasitic ances- the killing behaviour of the second instar. This be- tor, the question arose – what might be the reason haviour is probably connected with the missing for such a differentiation? We are convinced that host egg or young larva. Selective pressure on the it is caused by host antistrategy against its nest first instar depends more on intraspecific conflict parasite. There are several host behavioural fea- than on the elimination of the host. The early de- tures that are almost surely connected with these veloped second instar possesses mandibles al- coevolutionary effects. First, egg-laying under- most two times larger than those of the first instar neath the provision is interesting and unique be- and thus the survival probability of the first in- haviour of emphorine bees (subtribe Emphorina stars should be low in unequal combat. Killing only) (Michener 2000). It may represent an adap- behaviour of second instars is quite rare in bees tive answer to the behaviour of their nest parasite, and is often caused by the fact that first instars are which was destroying their eggs inside. Such be- inactive and remain surrounded by the egg haviour is currently known in Protosiris,which chorion (Torchio 1989, Garófalo & Rozen 2001). destroys host eggs with its sting. It was also ob- In megachilid bees there are some known species served that no first instar was alive in the cell of that are hospicidal in higher instars (Torchio Protosiris, even if the cell was repeatedly parasit- 1989, Rozen & Özbek 2004). Epeoloides first ised (Rozen et al. 2006). In such cases the defend- instar possess typical killing morphology, thus ing behaviour, like emphorine bees have, should the killing behaviour of the first instar is not out of be successful. Laying the egg beneath the provi- the question. sion clearly evolved in a common ancestor of the The short-term duration of the first instar subtribe Emphorina, because all genera possess Epeoloides is in high contrast with the long-term this character. We must think of an historical nest first instar of Leiopodus singularis and L. parasite of emphorine bees and a possibly ances- trochantericus.TheL. singularis egg incubation tor of current Leiopodus, which according to period is equalled the duration of the host’s cell cladistic analysis, probably arose from some provisioning (often partly) plus egg incubation osirine bee similar in parasitic behaviour to Pro- period plus the duration of at least the first two tosiris. larval stadia and usually the length of the third Analogously, the Macropis female staying at stadium as well (Rozen et al. 1978). On the other the entrance (and probably guarding it) during hand, L. lacertinus does not display such an ex- warm days may be the most important event that treme form of behaviour (Roig-Alsina & Rozen affects the Epeoloides female in its mode of para- 1994). The only found first instar of L. tro- sitism. The was clearly unsuccessful chantericus fed on second or third instar of when it tried to attack the guarded nest. We found Diadasina, so that its hospicidal behaviour could one nest with five cells and all of them contained be similar or the same as in L. singularis, with late cocoons of Macropis, even if several Epeoloides eclosion and killing the quiescent host larva females frequently visited this nest. When a closely before molting to the third or higher Macropis female left its nest during the most instar. risky time period (it also means if a female was ar- Both described species of larvae differ much tificially eliminated) the nest was always success- from each other both in morphology and behav- fully parasitised. This leads to the conclusion that iour. They may be different because they have this host behaviour is a very important strategy different life strategies, and, due to selective pres- against Epeoloides. Epeoloides must elude this sure, they changed a lot from the common antistrategy by changing its mode of parasitism. plesiomorphic growth plan. This is surely right The parasitic female could return to the nest later especially for mature larva of Leiopodus,which (1) after closing of the cell, or earlier, (2) before is superficially in the least degree dissimilar to the host female begins to guard the nest. If it any other bee larvae. After all, both tribes (Osirini comes to the nest later (1), the available cell and Protepeolini) may be closely related as the would be closed or closed and parasitised by the cladistic analysis of mature larvae suggests female that came first (2). Epeoloides has adap- (Straka & Bogusch 2007). Even though, they tively shortened the course of the first instar and 254 Straka & Bogusch • ENTOMOL. FENNICA Vol. 18 for this reason, the brood of the second visitor (1) Roig-Alsina, A. & Michener, C. D. 1993: Studies of the of the nest cannot be successful in intraspecific phylogeny and classification of long-tongued bees (Hymenoptera: Apoidea). — Univ. Kan. Sci. Bul. 55: conflict. 124–162. Roig-Alsina, A. & Rozen, J. G. Jr. 1994: Revision of the Acknowledgements. We would like to thank Jerome G. cleptoparasitic bee tribe Protepeolini, including biolo- Rozen and anonymous referees for their critical review and gies and immature stages. — Am. Mus. Nov. 3099: 1– improvement of this manuscript. We are also grateful to 27. Nick Stuart for checking English. The project and the pub- Rozen, J. G. Jr. 1991: Evolution of cleptoparasitism in lication were supported by the Research Programs MSM Anthophorid bees as revealed by their mode of parasit- No. 0021620828 and CTS MSM No. 0021620845. ism and first instars (Hymenoptera: Apoidea). — Am. Mus. Nov. 3029: 1–36. Rozen, J. G. Jr. 2000: Systematic and geographic distribu- tion of neotropical cleptoparasitic bees, with notes on References their modes of parasitism. — Anais do IV encontro sobre abelhas 2000, Ribeirão Preto — SP,Brasil: 204– Alves-dos-Santos, I., Melo, G. A. R. & Rozen, J. G. Jr. 210. 2002: Biology and immature stages of the bee tribe Rozen, J. G. Jr. 2001: A taxonomic key to mature larvae of (Hymenoptera: Apidae). — Am. Mus. cleptoparasitic bees (Hymenoptera: Apoidea). — Am. Nov. 3377: 1–45. Mus. Nov. 3309: 1–27. Bogusch, P. 2003: Hosts, foraging behaviour and distribu- Rozen, J. G. Jr. & Jacobson, N. R. 1980: Biology and im- tion of six species of cleptoparasitic bees of the mature stages of Macropis nuda, including compari- subfamily Anthophorinae (Hymenoptera: Apidae). — sons to related bees (Apoidea, ). — Am. Acta Soc. Zool. Bohem. 67: 65–70. Mus. Nov. 2702: 1–11. Bogusch, P. 2005: Biology of the cleptoparasitic bee Rozen, J. G. Jr. & Özbek, H. 2004: Immature stages of the Epeoloides coecutiens (Hymenoptera: Apidae: Osi- cleptoparasitic bee Dioxys cincta (Apoidea: rini). — J. Kans. Entomol. Soc. 78: 1–12. Megachilidae: : Dioxyini). — Am. Mus. Garófalo, C. A. & Rozen, J. G. Jr. 2001: Parasitic behav- Nov. 3443: 1–12. iour of with descriptions of its Rozen, J. G. Jr., Eickwort, K. R. & Eickwort, G. C. 1978: mature oocyte and larval instars (Hymenoptera: Api- The bionomics and immature stages of the clepto- dae: ). — Am. Mus. Nov. 3349: 1–26. parasitic bee genus Protepeolus (Anthophoridae, No- McGinley, R. J. 1981: Systematics of the Colletidae based madinae). — Am. Mus. Nov. 2640: 1–24. on mature larvae with phenetic analysis of apoid lar- Rozen,J.G.Jr.,Melo,G.A.R.,Aguiar,A.J.C.&Alves- vae. — Univ. Calif. Publ. Entomol. 91: i–xvi 1–307. dos-Santos, I. 2006: Nesting biologies and immature Michelette, E., Camargo, J. M. F. & Rozen, J. G. Jr. 2000: stages of the Tapinotaspidine bee genera Monoeca and Biology of the bee Canephorula apiformis and its Lanthanomelissa and of their Osirine cleptoparasites cleptoparasite Melectoides bellus: Nesting habits, flo- Protosiris and Parepeolus (Hymenoptera: Apidae: ral preferences, and mature larvae (Hymenoptera, Apinae). — Am. Mus. Nov. 3501: 1–60. Apidae). — Am. Mus. Nov. 3308: 1–23. Straka, J. & Bogusch, P.2007: Phylogeny of the bees of the Michel-Salzat, A., Cameron, S. A. & Oliveira, M. L. 2004: family Apidae based on larval characters with focus on Phylogeny of the orchid bees (Hymenoptera: Apinae: the origin of cleptoparasitism (Hymenoptera: Api- Euglossini): DNA and morphology yield equivalent formes). — Syst. Entomol. 32: 700–711. patterns. — Mol. Phyl. Evol. 32: 309–323. Švácha, P. & Danilevsky, M. L. 1987: Cerambycoid larvae Michener, C. D. 2000: The bees of the world. — The Johns of Europe and Soviet Union. Part I. — Acta Univ. Hopkins University Press, Baltimore and London. 913 Carol. — Biol. 30: 1–176. pp. Torchio, P. F. 1989: Biology, immature development, and Michener, C. D. & Greenberg, L. 1980: Ctenoplectridae adaptive behavior of montana, a cleptoparasite and the origin of long-tongued bees. — Zool. J. Lin. of Osmia (Hymenoptera: Megachilidae). — Ann. Soc. 69: 183–203. Entomol. Soc. Am. 82: 616–632.