Functional Morphology of the Male Genitalia and Copulation in Lower

AZO094.fm Page 331 Wednesday, September 5, 2001 10:03 AM

Acta Zoologica (Stockholm) 82: 331–349 (October 2001)

BlackwellFunctional Science Ltd morphology of the male genitalia and copulation in lower Hymenoptera, with special emphasis on the Tenthredinoidea s. str. (Insecta, Hymenoptera, ‘Symphyta’) Susanne Schulmeister

Abstract Institut für Zoologie und Anthropologie, Schulmeister, S. 2001. Functional morphology of the male genitalia and Abteilung Systematik, Morphologie und copulation in lower Hymenoptera, with special emphasis on the Tenthredinoidea Evolutionsbiologie, Berliner Str. 28, s. str. (Insecta, Hymenoptera, ‘Symphyta’). — Acta Zoologica (Stockholm) 82: D–37073 Göttingen, Germany 331–349

Keywords: A general description of the male reproductive organs of lower Hymenoptera Hymenoptera, Tenthredinoidea, is given. The terminology of the male genitalia is revised. The male external morphology, male genitalia, copulation genitalia of Tenthredo campestris are treated in detail as a specific example of the morphology. The interaction of the male and female parts during copula- Accepted for publication: tion is described for Aglaostigma lichtwardti. The possible function of sclerites 5 April 2001 and muscles of the male copulatory organ of Tenthredinoidea s. str. is discussed. Additional observations on morphology and function made in non- tenthredinoid lower Hymenoptera are included. The assumption that the gonomaculae act as suction cups is confirmed for the first time. The evolution of obligate and facultative strophandry is discussed. The stem species of all Hymenoptera was probably orthandrous and facultatively strophandrous. Susanne Schulmeister. Institute of Zoology, Berliner Str. 28, D–37073 Göttingen, Germany. E-mail: [email protected]

tionships among Hymenoptera. This scarcity of male genital Introduction characters is due to the fact that a comparative analysis in a The lower Hymenoptera known as ‘Symphyta’ or sawflies phylogenetic context does not exist. The detailed investiga- are a paraphyletic grouping lacking the wasp-waist of the tion of the morphology and function of the male copulatory Apocrita. Among them, the Tenthredinoidea s. str. [Argidae, organ of the Tenthredinoidea s. str. presented here was Cimbicidae, Diprionidae, Pergidae (= Pterygophoridae) and undertaken to serve as the basis for such an analysis. Tenthredinidae] are probably monophyletic (Vilhelmsen To date, the most extensive publication on the male exter- 1997, 2001). With about 7400 described species, the nal genitalia of ‘Symphyta’, especially their musculature, is Tenthredinoidea s. str. comprise the majority of the approx- that of Boulangé (1924). Other noteworthy papers on this imately 8000 described sawfly species (Goulet and Huber topic are those of Birket-Smith (1981), Crampton (1919), 1993). Peck (1937), Ross (1945), Michener (1956), E. L. Smith Since the ‘Symphyta’ are the basal groups of Hymeno- (1969, 1970a, 1970b, 1972), and Snodgrass (1941). The ptera, they play an important role in the elucidation of the copulation of sawflies is discussed in Boulangé (1924), relationships between holometabolous insect groups. Many Rohwer (1915), d’Rozario (1940), and E. L. Smith (1970a). morphological features of lower Hymenoptera have already The function of the genital muscles of a male braconid been studied intensively and used for phylogenetic analysis (Apocrita) was studied by Alam (1952). However, the present (Königsmann 1976, 1977; Vilhelmsen 1997, 2000; refer- paper treats the morphology and especially the function in ences therein), but very few characters of the male external much more detail. genitalia were employed in these studies. Of the five male There has been a lot of confusion about the terminology genital characters included in Vilhelmsen (2001), two were of the parts of the male genital organ in Hymenoptera. A coded as invariant in the Hymenoptera, and therefore only main source for this confusion is the fact that early authors three were potentially informative for the phylogenetic rela- studied mainly Apoidea, in which the male copulatory organ

AZO094.fm Page 332 Wednesday, September 5, 2001 10:03 AM

Male genitalia and copulation in Hymenoptera • Schulmeister Acta Zoologica (Stockholm) 82: 331–349 (October 2001)

is highly derived, and later on their terms were often applied exemplars from each species were examined. The male to non-homologous parts in other Hymenoptera. A treatment genitalia were dissected under a Zeiss stereomicroscope of all terms that have been applied to the male external geni- Stemi SV 6 (maximum magnification 50). Viewing the talia of Hymenoptera is beyond the scope of this paper, but objects under light coming from the side (obtained by point- is in preparation. However, the usage of the most important ing the tip of a goose-necked lamp at the side of the prepara- of the many existing terms is discussed in the present paper. tion dish) proved sufficient for a discrimination of the parts, An interesting feature of the male genital organ of Xyelinae so that staining was not necessary. Drawings were made with and Tenthredinoidea s. str. is that it is revolved as a whole by the aid of a camera lucida. 180° on its median axis (Crampton 1919). The normal con- Sclerites and membranes of the male genitalia can be hard dition is called orthandrous (orthandrious), the rotated con- to distinguish at the beginning. It proved useful to dry a gen- dition strophandr(i)ous (Crampton 1919). In strophandrous ital organ in a critical point dryer, apply a gold-coating (as for sawflies, the true ventral side becomes the apparent dorsal electron microscopy), and examine it under a stereomicro- side of the genital organ, which Boulangé (1924) termed ex- scope. Sclerites treated in this manner appear smooth and ventral. In this paper, the terms ventral and dorsal always shining, whereas membranes appear rough and dull. refer to the original (i.e. orthandrous) condition, i.e. ‘ventral’ Copulations were either observed (rather coincidentally) always means ‘anatomically ventral’. Strophandry probably in the field (in C. pygmeus, E. koehleri and T. temula) or initi- evolved independently in Xyelinae and Tenthredinoidea ated by putting a female and a male together in a vial (in s. str. (Vilhelmsen 1997, 2001). According to Rasnitsyn (1969; C. pygmeus, A. lichtwardti and M. ferruginea). Pairs fixed in in Ronquist et al. 1999), the torsion is already present in the copula were obtained by drowning them in alcohol (70–95%) pupa in Tenthredinoidea s. str., while in Xyelinae it appears during copulation. The pairs stayed coupled in about 30– during eclosion. No anatomical differences in the torsion of 50% of the attempts. Four pairs in copula of A. lichtwardti the male genitalia could be observed in these two taxa. were dissected. However, in the copulation of orthandrous as well as Aglaostigma lichtwardti was chosen for a detailed study of strophandrous Hymenoptera the ventral side of the male copulation because it was the only tenthredinoid species of genitalia is always facing up, whereas the dorsal side is always which several pairs in copula could be obtained. However, facing down, towards the substrate. In orthandrous species the drawings in the morphological part of this paper show this is achieved, for example, by curving the abdomen so that mainly Tenthredo campestris, because its male copulatory the tip of the abdomen comes to lie upside down. In organ is less derived than that of Aglaostigma and therefore strophandrous species such a measure is unnecessary facilitates comparison with other species. However, the dif- because the male genitalia already are upside down. ferences are not substantial, so that the general conclu- sions drawn about the functional morphology in the present paper should apply to both these species and most other Material and Methods Tenthredinoidea s. str. as well. (The differences between The species studied are shown in Table 1. these two species can be deduced to some extent from All specimens were fixed in Bouin’s fluid and kept in Fig. 2A,B vs. 4B,C; Fig. 8A vs. 8D; Fig. 7F vs. 9 A, and from ethanol (70%) until the preparation. If possible, two or more Fig. 6)

Table 1 Species studied in this paper Xyeloidea: Xyelidae: Macroxyela ferruginea (Say, 1824)

Tenthredinoidea: Blasticotomidae Runaria reducta (Malaise, 1931) Tenthredinidae: Nematus abbotii (Kirby, 1882) Tenthredo campestris (Linnaeus, 1758) Tenthredo temula (Scopoli, 1763) Macrophya annulata (Geoffroy, 1785) Aglaostigma lichtwardti (Konow, 1892) Elinora koehleri (Klug, 1817) Argidae: Arge rosae (Linnaeus, 1758) Arge rustica (Linnaeus, 1758) Pergidae: Lophyrotoma analis (Costa, 1864) Diprionidae: Macrodiprion nemoralis (Enslin, 1917)

Pamphilioidea: Megalodontesidae: Megalodontes cephalotes (Fabricius, 1781) = klugi (Leach, 1817) = spissicornis (Klug, 1824) Pamphiliidae: Cephalcia sp.

Cephoidea: Cephidae: Cephus pygmeus (Linnaeus, 1767)

Siricoidea: Siricidae: Xeris spectrum (Linnaeus, 1758) Anaxyelidae: Syntexis libocedrii Rohwer, 1915

AZO094.fm Page 333 Wednesday, September 5, 2001 10:03 AM

Acta Zoologica (Stockholm) 82: 331–349 (October 2001) Schulmeister • Male genitalia and copulation in Hymenoptera

Fig. 1—Male reproductive organs. —A. Schematic drawing to show the relations of the membranous parts of the integument, the sclerites and the inner reproductive organs. —B. One half of the male genitalia of Tenthredo campestris, looking onto the medio-sagittal plane (cf. Fig 7E,F). Scale in all ﬁgures: 1 mm.

deferens and is coiled into a spiral or lump before fusing with Results the side of the glandula mucosa. The latter can be described The male reproductive organs consist of inner and outer as a blind sac which can assume different shapes. An extreme parts. The latter have been called genitalia, external genitalia, example is a striking apomorphy observed in Megalodontes genital appendages, genital armature, armatura genitalis, cephalotes which has not been described until now. In this genital capsule, phallus, genital or copulatory organ, genital species, the glandula mucosa has three blind ends instead of or copulatory or phallic apparatus. The genital capsule is one, and the vesicula seminalis is ‘nestling’ among these situated between the ninth and tenth sternum, but due to dis- three ends. It would be interesting to examine other species placement it does not lie ventrally but rather at the tip of the of the Megalodontesidae in order to find out whether it is an abdomen concealed in the genital chamber, that is above the autapomorphy of the whole group. The caudal ends of the hypopygidium (ninth sternum) and below the proctiger (anal glandulae mucosae continue into the ducti ejaculatorii, papilla). The ninth sternum often has an apophysis, called which fuse to become the unpaired ductus ejaculatorius the spiculum, at its anterior tip. before entering the external genitalia. In Hymenoptera, the male copulatory organ is connected to the rest of the abdomen only via a thin membrane, the Membranous parts of the external male genitalia ducti ejaculatorii, muscles, tracheae and nerves; it can easily be pulled out and removed. It can also be easily turned Figure 1A shows the relationships between the sclerotized around. In strophandrous sawflies, Xyelinae and Tenthredi- and the membranous parts of the integument of the external noidea s. str., the rotated condition of the male genitalia is male genitalia, which is continuous with the internal ducts. permanent. In the past, there has been some confusion as to The ductus ejaculatorius passes between the two penisvalvae the condition in the Blasticotomidae, but work by Togashi (Fig. 1B cf. Fig. 4B,D) before being enlarged to constitute (1970) and Vilhelmsen (1997) and my own observation (in the endophallus, which is a more or less sac-like structure in Runaria reducta Malaise) show that at least Blasticotoma the centre of the genital organ (Fig. 1A,B). The transition filiceti Klug and Runaria reducta are orthandrous. from the ductus ejaculatorius to the endophallus can be formed in a way that produces the incorrect impression that they were separate structures and that the ductus penetrated Inner reproductive organs the wall of the endophallus. This transition is called the orig- The inner reproductive organs consist of the testes, vasa inal or primary gonopore. However, the ductus ejaculatorius deferentia, vesiculae seminales, glandulae mucosae (accessory is in no way separated from the endophallous, just as the glands) and ducti ejaculatorii (Fig. 1A). The vasa deferentia endophallic membrane is continuous with the ectophallic lead from the testes to the glandulae mucosae. The vesicula membrane which constitutes the surface of the copulatory seminalis is basically the enlarged proximal part of the vas organ. The opening of the endophallus to the outside is called

AZO094.fm Page 334 Wednesday, September 5, 2001 10:03 AM

Male genitalia and copulation in Hymenoptera • Schulmeister Acta Zoologica (Stockholm) 82: 331–349 (October 2001)

Fig. 2—Male external genitalia of Tenthredo campestris. ventral. —F. Penisvalva. —G. Volsella, outer face. —H. Volsella, —A. (Morphologically) dorsal side (which is actually ventral, due to inner face. —I. Parossiculus, distivolsella dotted. —J. Parossiculus, the strophandry). —B. Ventral side. —C. Cupula, seen from inside. distivolsellar apodema dotted. —D. Gonostipites and one harpe, dorsal view. —E. Same, from

AZO094.fm Page 335 Wednesday, September 5, 2001 10:03 AM

Acta Zoologica (Stockholm) 82: 331–349 (October 2001) Schulmeister • Male genitalia and copulation in Hymenoptera

the phallotrema, secondary gonopore, or gonotrema. The Latimere. The latimere (new term) consists of the gonostipes phallotrema is on the ventral side of the genital apparatus and (see discussion below) and the harpe (Crampton 1919) closed to a slit (Fig. 2B). Snodgrass (1941) wrote that the (Fig. 2D,E). In Cephidae, Orussidae and many Apocrita, the endophallus ‘Probably ( ... ) is always eversible’, but I could ‘outer clasper’ of the male external genitalia is represented only observe this in Cephidae. only by a single piece. For this piece Peck (1937) proposed The term penis has probably been applied to non- to employ a term of its own – for which he chose gonoforceps homologous structures in the different insect orders. In – because we cannot tell for certain whether this single piece Hymenoptera, the membrane situated between the distal evolved through a fusion of gonostipes and harpe and is, heads of the penisvalvae, i.e. the valvicepes, was sometimes hence, homologous to the latimere or through a total reduc- termed penis (e.g. Michener 1956). Some authors (e.g. tion of the harpe, thus being homologous to the gonostipes Snodgrass 1941) call this distal portion of the endophallic alone. Some authors claimed that there was a line on the membrane a penis only if it forms a structure separate from gonoforceps in certain species which is taken as evidence of the penisvalvae, as is the case in Apoidea. In ‘Symphyta’ there a fusion, but even if there were such a line it could still have is no structure to which this second definition would apply. originated secondarily. The word aedoeagus is usually understood as meaning the Until now, the word harpes has not only been used for the intermittent part of the genitalia, which in Hymenoptera plural, but erroneously for the singular as well (e.g. by Ross are the valvicepes and the membrane between them. For 1945; Wong 1963; Königsmann 1976). the definition of phallus see Snodgrass (1941; p. 3). In the The term stipes was applied to a part of the male external Hymenoptera, the phallus is the entire male genital capsule. genitalia of Hymenoptera for the first time by Thomson (1871/72). He used it for Bombus in which the harpe is reduced and the volsella merely a small scale on the medial Sclerites of the male genitalia face of the gonostipes (according to Snodgrass 1941), which In Hymenoptera, the male copulatory organ consists of four is not depicted in Thomson’s figure. Therefore it is impos- main sclerites which are connected only through membrane sible to say which parts Thomson understood as stipes. and muscles (unless secondarily fused): an unpaired basal Crampton (1919) changed Thomson’s term to ‘gonostipes’ ring, two pairs of clasping organs serving to grip the female and applied it to what is today called gonostipes and volsella. during copulation, and a pair of median skeletal supports for Later on, Crampton’s term ‘gonostipes’ has been mistakenly the intromittant organ. These main sclerites are described in used to stand for the gonostipes alone, without the volsella detail below. (e.g. Peck 1937; Ross 1945; Königsmann 1976). But since An overview of the most important terms for sclerites of gonostipes is a well-known term and as today’s meaning has the male genitalia and their different usages is given in Fig. 3. consistently been applied, I do not want to change this and Tables like this and lists of synonyms are numerous in the use it in its modern sense, i.e. excluding volsella. I want to literature (e.g. Boulangé 1924; Beck 1933; Snodgrass 1941; point out that the term should not imply any homology to Michener 1956; E. L. Smith 1970a,b; Birket-Smith 1981; any parts of mouthparts or legs. Kopelke 1982), but none of them is entirely correct. Birket-Smith (1981) suggested the use of stipes instead of gonostipes (according to the rule of priority), but used the Cupula. The proximal basis of the genital organ is surrounded term stipes for gonoforceps/latimere, i.e. differently from by the unpaired cupula [coined by Birket-Smith (1981) after either the original or the modern sense of gonostipes. Since cupule (Audouin 1821); = basal ring (Crampton 1919)] this would only add confusion and because the term stipes (Fig. 2A–C). already exists for a mouthpart (which is why Crampton Michener (1944a,b, 1956) called the cupula gonobase, replaced it with gonostipes), I reject the use of the word but ‘gonobasis’ is also being used for non-homologous parts stipes. I also do not follow Birket-Smith’s (1981) suggestion in basal insects (e.g. Willmann 1998). to use the term harpide (Audouin 1821) instead of harpe In the primitive condition, the cupula is a closed ring because Audouin coined the term harpide for a genital which is broad dorsally, but narrow on the ventral side, where part of the bumblebees, which do not have a harpe. it is medially protruded into an apophysis, the gonocondyle Therefore the harpide cannot be homologous to the part (Crampton 1919). The ventral side of the cupula can be called harpe. strengthened by an internal ridge (Fig. 2C) or thickening. Gonostipes and harpe have also been called gonocoxite The cupula surrounds the basal opening termed foramen and gonostylus (Michener 1944a, 1944b, 1956). The terms genitale (Snodgrass 1941). It can be fused in part with the imply a homology with coxa and stylus, and I prefer to gonostipites, e.g. in Arge. The dorsal portion can be reduced employ neutral terms for the time being. Furthermore, if to a narrow string as in Aglaostigma lichtwardti (Fig. 4B) or parts of the hymenopterous male genitalia are homologous to completely reduced as in Arge (Fig. 4A) and Megalodontes the gonocoxa, this would probably be the cupula, gonostipes cephalotes. The cupula can also be missing completely as in and volsella together and not just the gonostipes. Regarding Lophyrotoma analis. the stylus, there was disagreement whether this would be the

AZO094.fm Page 336 Wednesday, September 5, 2001 10:03 AM

Male genitalia and copulation in Hymenoptera • Schulmeister Acta Zoologica (Stockholm) 82: 331–349 (October 2001)

Fig. 3—Some terms of the male genitalia of Hymenoptera as used – but not necessarily coined – by some authors. Apostrophes indicate different uses of the same term.

harpe (e.g. E. L. Smith 1969, 1970a,b) or the digitus of the lateral larval phallomeres (= ‘parameres’). Moreover, (Birket-Smith 1981). if the cupula is a derivative of the gonostipites – and not a The popular term paramere is highly ambiguous, even if newly formed sclerotization of the basal ectophallic mem- only its application to the hymenopterous genitalia is con- brane, as Snodgrass assumed – the term paramere would also sidered: in Fig. 3 there are just two of many different usages. have to comprise one half of the cupula. (The same problem Verhoeff (1893) applied the term paramere to gonostipes, is inherent to the term ‘basimere’.) Most contemporary harpes and volsella. Beck (1933) and Peck (1937) employed hymenopterists (except for Königsmann 1976) still use the it for the penisvalva. Snodgrass used it ﬁrst for the harpe only word paramere in the sense of Snodgrass (1941), obviously (Snodgrass 1941) and later on for gonostipes and harpe not realizing that Snodgrass (1957) himself objected to this together (Snodgrass 1957), arguing that both are derivatives with good reason. Because the term paramere is highly

AZO094.fm Page 337 Wednesday, September 5, 2001 10:03 AM

Acta Zoologica (Stockholm) 82: 331–349 (October 2001) Schulmeister • Male genitalia and copulation in Hymenoptera

Fig. 4—Male external genitalia of —A. Arge rustica (morphologically) dorsal side —B. Aglaostigma lichtwardti, dorsal —C. same, ventral, and —D. Macrodiprion nemoralis, dorsal.

ambiguous and used incorrectly in adult Hymenoptera, I Apocritans corresponds to the original harpe and can there- prefer not to use it at all. fore be designated with the same name. But the secondary The discussion about the terms paramere sensu Snodgrass division itself is not identical to the primary division. (1941) (= harpe) gets even more complicated when consider- Moreover, secondary divisions probably evolved several ing the Apocrita. The stem species of the Apocrita probably times independently in the Apocrita. To make clear that the had a gonoforceps, i.e. no or no separate harpe. In some appendices of some Apocrita are not directly derived from Apocrita, however, the gonoforceps is secondarily divided the primary appendices (= harpes) of some lower Hyme- into two parts (Snodgrass 1941; Königsmann 1976). If we noptera, one should consider naming each independently assume that the gonoforceps developed through fusion of derived secondary appendix differently. For this reason, gonostipes and harpe, so that the distal portion of the gono- Birket-Smith (1981) proposed the term kleisiades for the forceps corresponded to the harpe, one could assume – as secondary appendix in Lapidus and Dorylus (Dorylidae, Snodgrass (1941) did – that the secondary appendix in some Formicoidea).

AZO094.fm Page 338 Wednesday, September 5, 2001 10:03 AM

Male genitalia and copulation in Hymenoptera • Schulmeister Acta Zoologica (Stockholm) 82: 331–349 (October 2001)

The gonostipites are the central and largest sclerites of Anaxyelidae, Orussidae and Apocrita. The absence of the genital organ and constitute its main framework. The two gonomaculae in the only extant anaxyelid Syntexis libocedrii gonostipites abut dorsally as well as ventrally so that they (Middlekauff 1964) was confirmed in the present study. surround the genital apparatus (Fig. 2D,E). The medial area Although the gonomacula is coded as absent in the Blastico- of the dorsal side of the gonostipes is called the parapenis tomidae Runaria reducta, Paremphytus flavipes (Takeuchi) and (Crampton 1919). In some Tenthredinoidea, the parapenis is Blasticotoma filiceti in Vilhelmsen (in press), I found a tiny set off against the rest of the gonostipes (Fig. 4B), sometimes rest of a gonomacula in the blasticotomid Runaria reducta, only connected to it by a narrow bridge (Figs 2D, 4D). but was not able to study any other species of that family. Boulangé (1924) mentioned it as ‘le pont reliant le para- Since a muscle inserts on the gonomacula, several authors penis au flanc ...’ The connection between the parapenis assumed that it probably functions like a suction cup. So far, and the rest of the gonostipes, be it a narrow bridge or a this theory had not been confirmed. rather abstract cranial-caudal line (as in Fig. 4B), has been termed the parapenisjugum in the present study. Volsella. On the ventral side of the male external genitalia, In the primitive condition, the parapenes touch each between the gonostipes and the penisvalva, lies the volsella other, but are not fused (Figs 2D, 4B). In some species, how- (sensu Snodgrass) (Fig. 2G–J). It is a pincer-like structure ever, a proximal fusion occurs secondarily (Fig. 4A,D). and consists of two separate sclerites, the lateral parossiculus Crampton (1919) only employed the term parapenis when and the smaller gonossiculus (both Crampton 1919) = this structure was clearly set off. Boulangé (1924; p. 59) digitus (volsellaris) (Snodgrass 1941). Both sclerites are argued that the insertion site of muscle j should be called attached to each other along their middle parts by a narrow parapenis, even if not set off against the rest of the gono- unpigmented, but sclerotized line. In some Hymenoptera stipes. But since the insertion site of a muscle can wander to (e.g. Cephalcia sp., Cephus pygmeus and Xeris spectrum), the a different place during the course of evolution (see below), volsella is fused secondarily with the gonostipes. it cannot be decisive for naming a sclerite. Therefore, the Dufour (1841) coined the term volselle after the Latin parapenis is here defined as the baso-medial part of the dorsal word volsella and marked it in his drawings of the male gonostipes. In all but one of the species I examined, this area genitalia of some Apocrita. His definition and figures are is the insertion site of muscle j; only in Megalodontes cepha- quite vague, but it seems that he applied ‘volselle’ to differ- lotes, did muscle j insert on an inflection of the medial edge ent structures, namely the penisvalva, the volsella sensu of the dorsal gonostipes; in this case it is uncertain which part Snodgrass, and the ventral lamella of the gonoforceps. Peck is homologous to the parapenis. (1937), Snodgrass (1941), Ross (1945), Alam (1952), The basal edge of the gonostipes (including the parapenis) Michener (1944a,b, 1956), Königsmann (1976), Gauld and is reinforced by an inflection (Fig. 2D). Where the parapenis Bolton (1988), Schedl (1991), and Ohl (1996) – among is set off against the rest of the gonostipes, its lateral edge is others – called the gonossiculus and parossiculus together reinforced by a thickening (Fig. 2E). On the ventral side, the the volsella. Although Birket-Smith (1981) wanted ‘to use median edge of the gonostipes does not always form a dis- only terms that never have been used for more than one tinct line, but may instead turn gradually into membrane. structure in the Hymenoptera’, he employed the word Basally, the gonostipes is elongated into a brace-like struc- volsella in the sense of Crampton (1919), i.e. for the paros- ture, the gonostipital arm (Crampton 1919) (Fig. 2D,E). siculus only, obviously not realizing that contemporary However, it is not possible to draw a clear line between the authors were using it differently and that the term had not been gonostipal arm and the rest of the gonostipes (except maybe coined by Crampton himself. Crampton (1919), Boulangé in extreme cases as in the Pamphilioidea and Xiphydriidae). (1924), and Birket-Smith (1981) did not have a specific term The tip of the gonostipital arm is an insertion site for muscles to summarize gon- and parossiculus. E. L. Smith (1969, h and i and is here termed the apex gonostipitis (new term) 1970a, 1970b, 1972) employed the expression ‘Section 3 of (Fig. 2E). gonocoxite lX’. Since there exists no good alternative and Boulangé (1924) called it apophyse principal (which has since nearly all authors used and still use the term volsella in been mistaken as being synonymous with the gonostipal the sense of Snodgrass (1941), I do not want to replace this arm), but it is not an apophysis sensu Snodgrass (1935). popular term in spite of its ambiguity. Contrary to the case Each harpe is formed like a hollow dish and situated at of the word paramere, there is no morphologically correct or the disto-lateral edge of a gonostipes. The harpes are covered wrong meaning for volsella. with many bristles (which are not depicted in the drawings The parossiculus consists of a basal ‘handle’, termed the presented in this paper). There is a membranous area called basivolsella, and a distal ‘hook’, the distivolsella (Fig. 2I), the gonomacula (Crampton 1919) = ventouse (french for part of which is a basally directed muscle insertion site called sucker; Dufour 1854) = cupping disk (Snodgrass 1941) on distivolsellar apodeme (Fig. 2J) (all three terms by Peck the distal tip of the harpe (Fig. 7C) in Xyelidae, Pamphilii- 1937). The distivolsella (including the apodeme) was termed dae, Megalodontesidae, Siricidae and Xiphydriidae. The cuspis (volsellaris) by Snodgrass (1941). It must be empha- gonomaculae are absent in Tenthredinoidea s. str., Cephidae, sized that the basivolsella and distivolsella (cuspis) are not

Acta Zoologica (Stockholm) 82: 331–349 (October 2001) Schulmeister • Male genitalia and copulation in Hymenoptera

separate in any way, they make up one contiguous sclerite. valvurae are long enough to pass through the foramen Cuspis and digitus are usually opposable and are used as a genitale. Where the valviceps connects to the valvura, there clasper. In summary, the volsella can be distinguished either is a small side branch, the ergot (french for spur; Boulangé in parossiculus and gonossiculus or in basivolsella, cuspis 1924), which serves as an insertion site for several muscles. (distivolsella) and digitus (gonossiculus). The ergot can be reduced, e.g. Tenthredo campestris; in this When used together, the terms basivolsella and dis- case the muscles simply insert on the valvura at the position tivolsella have the disadvantage that this combination seems where the ergot would have sat. to imply that together they form the entire volsella, which can In the Nematinae (Tenthredinidae), the penisvalvae show lead to two misinterpretations: Either, that the volsella was a derived, complicated structure, described by Ross (1945) identical with the parossiculus, without the gonossiculus = and Wong (1963). The fine structure and evolution of the digitus [which might have led Birket-Smith (1981) to use penisvalvae were treated extensively by E. L. Smith (1969, volsella in this way, contrary to all contemporary authors], or 1970a, 1972). to the misinterpretation that the distivolsella comprises not only the cuspis but also the digitus (e.g. in Schedl 1991). Other sclerites of the aedoeagus. The median sclerotized style Similarly, the fact that the terms cuspis and digitus are usu- (Ross 1937) = detached rhachies ( E. L. Smith 1969, 1970a,b) ally used together as a pair has similarly led people to believe = ventral rod of aedoeagus (Snodgrass 1941) is present in that they together make up the whole volsella, i.e. that the the Cephidae and Siricidae and, according to Smith (1970a), cuspis would be the entire parossiculus (e.g. in Kimsey and has arisen through splitting from the penisvalvae. It is a long, Bohart 1990; Kopelke 1982). thin sclerite and lies on the median axis of the ventral side of The basivolsella runs parallel to the median edge of the the external genitalia, between the two penisvalvae. The gonostipes and although parts of it can be hidden behind the median rod is situated in the place where in other groups gonostipes and membrane (Figs 2B, 4C), it is an external would be the phallotrema. In Cephidae, its basal end is fused structure. The carina volsellaris = volsellar ridge (both terms with the gonostipes, but in Siricidae it is not. Snodgrass 1941) = volsellar strut (Peck 1937) is an apodeme In Apoidea there are different sclerotizations of the dorsal running across the length of the basivolsella, forming a ridge penis membrane, e.g. median rod, spatha (see Snodgrass internally and a sulcus externally (Fig. 2G,H). The dis- 1941). tivolsella bears bristles (not depicted). According to the definitions of structural terms given Fibula ducti. There is a very small sclerite situated on the by Ronquist and Nordlander (1989), the carina volsellaris ducti ejaculatorii where they unite to form an unpaired duc- is a ridge and not a carina, since they define carina as an tus (Fig. 4A). In Arge rosae, for example, it is represented by ‘external, simple ridge’. Others, however, seem to regard a two small sclerotic plates situated ventrally and dorsally of carina also as an internal ridge (e.g. Gibson 1985). the ductus ejaculatorius, and dissection shows that there is a Ross (1945) wrote about the gonossiculus: ‘The gonola- minute sclerotized bridge connecting the dorsal and ventral cinia has an apical portion or apiceps (ap) and a basal pro- parts in the median plane. The one depicted in Fig. 4A is longation or basiura (ba).’ Unfortunately, it does not become unusually large; normally this sclerite is much smaller, often clear from his text or figures where he drew the line to these transparent and inside the ductus, so that it is difficult to see. parts. Therefore, I replace his terms with digiceps and To my knowledge, this sclerite has been depicted in a sawfly digiura and define the digiura as the free proximal handle of only by D. R. Smith (1990; p. 46), in Perreyiella, a pergid, but the gonossiculus, and the digiceps as the rest, which is con- not named, and I term it fibula ducti. nected to the parossiculus (Fig. 2G). In Lophyrotoma analis, the fibula ducti is unusually large. In Arge rosae, it is large as well, but transparent, while in Penisvalva. In the middle of the external genitalia, parallel to numerous other tenthredinoid and non-tenthredinoid the medio-sagittal plane, lie the two penisvalvae (Crampton ‘Symphyta’ there is often only a trace of the fibula ducti. In 1919), each of which has approximately the shape of a spoon other species, e.g. Tenthredo campestris, it is absent. (Fig. 2F). The apical ‘disc’ is called valviceps, the long Clausen (1938: 247 f.) described a similar sclerite situated ‘handle’ valvura (both terms by Ross 1945). Along its rim inside the unpaired ductus ejaculatorius of Formica rufa, (dorsally, apically and ventrally), the valviceps is continuous which he termed Sperrkeil. Further research has to be done with ectophallic membrane, by which it is connected to the to determine whether the fibula ducti and the Sperrkeil are other valviceps (Fig. 1A,B). The ventral membrane stretches homologous. The fibula ducti might even have a homologue dorsally to form a trough, the endophallus. In some ‘Symphyta’ in Mecoptera, the Ostialsklerit (Willmann 1981, 1989). The (e.g. Cephus pygmeus and Xeris spectrum) and in many Apo- Ostialsklerit is also situated at the end of the ductus ejacula- crita, the valvicepes are not separate dorsally. The valvura is torius, but – contrary to the fibula ducti and the Sperrkeil – not connected to any membrane, because it is an internal is furnished with a pair of muscles. All these sclerites seem apodeme of the valviceps extending into the lumen of the to support the assumption that the ductus ejaculatorius is of genital organ. In many species, e.g. Tenthredo campestris, the ectodermous origin (Snodgrass 1941: 41957 : 2).

Male genitalia and copulation in Hymenoptera • Schulmeister Acta Zoologica (Stockholm) 82: 331–349 (October 2001)

Fig. 5—Male and female Aglaostigma lichtwardti in copula. Male parts are dotted. Of the female, only the end of the abdomen is shown. Of the male, the end of the abdomen is shown in A. In B–D the male abdomen was removed so that only the genitalia are shown. —A. Lateral view. —B. Same, left half of female removed except for the seventh sternum. —C. Ventro-lateral view. —D. Ventral view.

Although the males of Macroxyela are orthandrous, I Copulation observed that pairs of Macroxyela always copulated in the Observations in live Tenthredinidae show that during copu- manner of strophandrous sawﬂies, i.e. end-to-end with both lation male and female of these strophandrous species are sexes standing on the substrate. both standing on the substrate facing opposite directions. Preparations of pairs of Aglaostigma lichtwardti killed dur- The male slips the tip of his abdomen under that of the ing copulation are pictured in Fig. 5. The harpes are pressed female and presses his harpes onto the sides of her seventh against the lateral sides of the seventh sternum of the female sternum. In Elinora koehleri and Tenthredo temula, I observed (Fig. 5A). The exerted pressure can be strong enough to dent that the male often has one or both of his hind legs on the the female’s sternum. The volsellae hold the seventh sternum female’s wings during copulation. The females sometimes of the female between cuspis and digitus (Fig. 5B). The move during copulation, dragging the male behind. penisvalvae are inserted into the female genital opening Pairs of the orthandrous species Cephus pygmeus were (Fig. 5B,C).The parapenes lie dorsally of the female’s observed to copulate with the male on the female’s back. seventh sternum, i.e. the sternum is ‘squeezed’ between the Some females kept feeding on pollen and after a while moved parapenes and the rest of the gonostipes (Fig. 5C,D). their saws against the male genitalia, seemingly trying to The assumption that the gonomaculae (which are not remove them. present in Tenthredinoidea s. l., Cephidae, Orussidae and

Acta Zoologica (Stockholm) 82: 331–349 (October 2001) Schulmeister • Male genitalia and copulation in Hymenoptera

Apocrita) can act as suction cups could be shown to be thredo campestris, muscle f is derived in that it inserts laterally correct in the present study. When a male and a female of on the cupula. Macroxyela ferruginea were put together in a vial, the male Each penisvalva is moved (mainly) by five pairs of muscles usually stretched his copulatory apparatus out of the genital which run to the gonostipes, four of which move the penis- chamber. During this stage, the gonomaculae of several males valva in the medio-sagittal plane (Fig. 7F). Muscles h and i became attached to the glass of the vial through adhesion. connect the penisvalva to the ventral apex gonostipitis, The males were then unable to pull away although they whereas muscles j and k insert on the dorsal parapenis; j on braced themselves against the glass, exerting all possible the main part of the inner face of the parapenis, k on its force. This could last for 1 or 2 minutes, during which a median edge. At the penisvalva, muscles h and j insert on the pulsating action of the gonomaculae was observed through apex of the valvura and i on the ergot (or the equivalent the glass of the vial under the stereomicroscope. In some place). k is a more or less fan-shaped muscle which extends cases, the male could be freed only by pulling it with a pair from the parapenis to the median side of the penisvalva of tweezers from the vial. (Figs 7F, 9A). The contraction of any one of these four muscles can have different consequences, depending on the state or action of the others. Of course, another important Morphology and function of the muscles of the external factor for the movements of the penisvalvae are the spatial male genitalia relations of the insertion sites, which can differ considerably Boulangé (1924) studied 46 species of ‘Symphyta’ and found between taxa. Figure 9 shows the male external genitalia of 25 pairs of muscles, which he named with lower case letters Aglaostigma lichtwardti in the normal position (Fig. 9A) from a to v, plus x, z and si. When Snodgrass (1941) cited and in copula with revolved penisvalva (Fig. 9B). Boulangé’s work, he changed the letters to numbers. Figure 6 Muscle l (Fig. 7G) begins at the ergot (or the equivalent compares the two nomenclatures and gives the insertion sites point on the penisvalva) and ends (dorso)laterally at the gono- of all muscles presently known in lower Hymenoptera. Not stipes. In species with a demarcated parapenis, the insertion all muscles are present in every species; Tenthredo campestris, site lies laterally of the parapenisjugum. The muscles l thus for example, has 18, Macroxyela ferruginea 21. pull the penisvalvae laterally, i.e. away from each other, prob- The genital organ is connected to the hypopygidium ably after insertion of the penisvalvae into the genital opening (ninth sternum) by three pairs of muscles (Fig. 7A). Muscle of the female. Since the lateral movement of the penisvalvae a runs from the spiculum to the gonocondyle. Muscle b is apparently more restricted proximally than distally, the begins at the spiculum and ends at the ventral side of the cup- contraction of the muscles l probably spreads them apart dis- ula. Muscle c inserts on the gonocondyle and laterally on the tally. This action presumably opens the phallotrema to make ninth sternum. These three muscles work to move the genital way for the sperm and anchors the male copulatory appara- capsule as a whole: muscle c pulls it distally, a and b act tus in the female. antagonistically. The action of muscle c can probably be Muscle m – not found in Tenthredinoidea s. str. – extends enhanced through the application of haemolymphic pressure from the apex of the valvura to the digiceps. Muscle n runs at the end of the abdomen (Kluge 1895; p. 189). The muscles from the apex of the valvura to the digiura (Fig. 7E). The b and c are presumably able to turn the genital capsule side- insertion sites of the muscles m and n are thus very close to ways as well. In strophandrous species, these three muscles each other. However, m and n differ by the course they take are twisted and crossed due to the inversion of the copulatory from one insertion site to the other: m lies laterally of the organ (Fig. 7A); the muscles b both pass the gonocondyle on penisvalva and of muscle i (and s, if present), whereas n runs the same side (cf. E. L. Smith 1972). medially of the penisvalva over most of its length, before Cupula and gonostipes are connected through up to four crossing the penisvalva to reach the digiura. In its middle, pairs of muscles. In the presumed plesiomorphic condition muscle n is often attached to the endophallic membrane, (e.g. in Macroxyela ferruginea, Fig. 7C,D), all four pairs are close to the primary gonopore and can then appear bipartite. present: d and e on the ventral side, f and g on the dorsal side. Due to this connection muscle n possibly serves to open or Muscles d and f begin medially at the cupula and end later- close the primary gonopore. Apart from this, the function of ally at the gonostipites. Both pairs of muscles pull the gono- muscles m and n is difficult to deduce since both of their stipites laterally, away from each other, so that they open up ends insert on movable elements. to be able to fit around the abdomen of the female. Muscles Two muscles connect the volsella, or rather the parossicu- e and g insert laterally on the cupula and medially on the gono- lus, to the gonostipes: o and p (Figs 8A–D, 9C). Muscle o stipites. They act in antagonism to muscles d and f to pull inserts on the lateral basal portion of the basivolsella and dis- the gonostipites towards each other, especially to grasp the tally on the gonostipes (near the harpe). Muscle p runs from female during copulation. In all examined Tenthredinoidea the distivolsella to the basal section of the gonostipes. I found s. str. except for Nematus abbotii, only two pairs of muscles that three muscles instead of two connect parossiculus and are present (Fig. 7B), and in species with a partly or totally gonostipes in the species Aglaostigma lichtwardti, Nematus reduced cupula there might be just one or none at all. In Ten- abbotii and Runaria reducta. Two of these three muscles seem

Male genitalia and copulation in Hymenoptera • Schulmeister Acta Zoologica (Stockholm) 82: 331–349 (October 2001)

Fig. 6—Genital musculature of male lower Hymenoptera with in the third column is left empty, the names of Boulangé have been special reference to Tenthredo campestris (T. c.) and Aglaostigma used. The insertion sites given for muscles d–g correspond to the lichtwardti (A. l.). The ﬁrst and second columns list the names of the plesiomorphic condition. In the sixth and seventh column is muscles introduced by Boulangé (1924) and Snodgrass (1941). The indicated which muscles are present (+) or lacking (–) in the species third column shows the changes made in the present paper; if the cell mentioned above.

Acta Zoologica (Stockholm) 82: 331–349 (October 2001) Schulmeister • Male genitalia and copulation in Hymenoptera

Fig. 7—Muscles of the male external genitalia of lower Hymenoptera d and e. —D. Dorsal side of external genitalia with muscles f and g. (part 1), shown for Tenthredo campestris and Macroxyela ferruginea. —E. One half of external reproductive organs, showing muscle n in (C, D) —A. Dorsal side of ninth sternum, ventral side of external the medio-sagittal plane. —F. Same, with muscle n removed, genitalia, and muscles a, b and c. —B. Inner face of cupula with showing muscles h, i, j, k and si. —G. Gonostipes (without muscles e and f. —C. Ventral side of external genitalia with muscles parapenis), harpe, and penisvalva with muscles l and i.

Male genitalia and copulation in Hymenoptera • Schulmeister Acta Zoologica (Stockholm) 82: 331–349 (October 2001)

Fig. 8—Muscles of the external male genitalia of lower Hymenoptera —E. Dissected male genitalia (ventral view) showing muscles i, n (part 2), shown for Tenthredo campestris and Aglaostigma lichtwardti and si. Some parts have been removed, including the distal part of (D). A–D show the volsella with muscles o, p and qr: — A, D, seen n on the left part of the ﬁgure. —F. Harpe and distal part of from the outside of the genitalia, —B, C. viewed from the inside. gonostipes with muscles t′′′, t′′′′′′ and u.

to have evolved through splitting of o; therefore I termed correspond to muscle o′′′ and not o. Many more species them o′′′ and o′′′′′′. Muscle o′′′ lies in the same position as o, would have to be studied in order to elucidate the evolution whereas o′′′′′′ inserts more distally on the basivolsella. of the muscles of the ‘o-complex.’ However, at present I cannot say with absolute certainty The muscle o′′′′′′ has not been described in ‘Symphyta’ until now, that o′′′ and o′′′′′′ have evolved through the splitting of o. More- but Snodgrass (1941) described a muscle (termed 20), which, over, even if this was the case, it is possible that some species according to his description, seems to have the same inser- which have only one muscle connecting basivolsella and gon- tion sites as muscle o′′′′′′. Snodgrass (1941) found this muscle ostipes (e.g. Tenthredo campestris), evolved from an ancestor only in Vespoidea. At present, it is not possible to make a having both o′′′ and o′′′′′′, so that this single muscle would statement as to whether muscle 20 is homologous to muscle o′′′′′′.

Acta Zoologica (Stockholm) 82: 331–349 (October 2001) Schulmeister • Male genitalia and copulation in Hymenoptera

q and r lie side by side and mostly appear as one muscle, so I cannot say that they should be considered as two muscles, just because the insertion site happens to be on both sides of the carina volsellaris in certain species. Snodgrass (1941) regarded q and r as one muscle, which he termed 18. I follow him in that I treat q and r as one muscle, qr. Muscle qr pulls the distivolsellar apodeme towards the basivolsella. Muscle s extends from the medial basal part of the basivolsella to the digiura. Its function is probably to pull the digiura basad to open the volsella. Muscle si starts next to s at the basivolsella and ends at the ergot or the equivalent spot of the penisvalva (Fig. 8E). The muscle si got its name because Boulangé (1924) found that its position is intermediate between the muscles s and i. Both s and si are present in Cephalcia sp. In Macroxyela ferruginea and Xeris spectrum, I found only muscle si. Boulangé (1924) claimed that both muscles are present in the Tenthredininae, but the Tenthred- ininae examined in this study had only one muscle. Macro- phya annulata and Aglaostigma lichtwardti lack muscle s, and in Tenthredo campestris there is one muscle which runs from the basivolsella to a membrane between the digiura and the penisvalva (Fig. 8E) and is therefore intermediate between s and si. At present, I cannot say whether this muscle is homologous to both s and si or only to one of these. In those ‘Symphyta’ in which a harpe is present, Boulangé (1924) found two muscles which connect the harpe to the gonostipes: t and u (Fig. 8F). My own studies show that t has the tendency to separate into two parts, t′′′ and t′′′′′′; but there seems to be some variation in the degree of separation of these parts even within one species. Muscle t′′′′′′ inserts laterally on the gonostipes and on the edge of the medial face of the harpe. Muscle t′′′ runs parallel to t′′′′′′, a bit more distally. Both muscles probably bend the harpe inwards. The fan-shaped muscle u begins at the distal edge of the gonostipes and ends with its fanned side in the harpe – sometimes on its medial face, sometimes on the lateral. Muscle u probably works as an antagonist —External male genitalia of Aglaostigma lichtwardti in normal Fig. 9 ′′′ ′′′′′′ position (A) and from a specimen killed during copulation to t and t . The present study showed that in some non- (B, C). See text for explanation. —A, B. one half of genitalia, tenthredinoid species (e.g. Macroxyela ferruginea, Megalodontes looking onto the medio-sagittal plane. —C. Ventral view. cephalotes and Cephalcia sp.) u is divided into two separate, fan-shaped muscles running parallel to each other. A gonomacula, if present, is worked by muscle vs. When In species in which the parossiculus is fused with the the gonomacula is pressed against a surface and muscle v con- gonostipes (e.g. Cephalcia sp., Cephus pygmeus and Xeris tracts at the same time, the gonomacula acts as a suction cup. spectrum), none of the o muscles is present. According to Boulangé (1924), there is an exceptional Muscles o, o′′′, o′′′′′′ and p move the volsella as a whole. Muscle muscle in Abia lonicerae L., named x, which connects the two p can also aid muscle qr to bend the distivolsella (see below), valvurae near their apices. provided that the parossiculus is secured in its position – In Cephus pygmeus, muscles z extend between the unpaired either by contraction of muscle o/o′′′ or by being immovably median rod and the two valvurae. fused with the gonostipes. That muscle p should have this second purpose is suggested by the fact that p – in contrast Discussion to o – is still present in Cephalcia (Pamphiliidae), in which the volsella is secondarily immovably fused with the gonostipes. Copulatory behaviour Muscles q and r both begin at the distivolsellar apodeme and end at the basal section of the basivolsella: q laterally, r Enslin (1912) and Boulangé (1924) noted that the tenthredi- medially of the carina volsellaris (Fig. 8A,C,D). Muscles noid male walks backwards towards the rear end of the

Male genitalia and copulation in Hymenoptera • Schulmeister Acta Zoologica (Stockholm) 82: 331–349 (October 2001)

female, slips the tip of his abdomen under hers and places his present) are then thrust into the vagina. Sperm is injected harpes on the sides of her seventh sternum. Gordh (1975) probably both by peristalic motion in the gonapophyses (the observed several copulations from laboratory-reared virgin pectines and setae would sweep the material distally) and specimens of Hemitaxonus dubitatus (Norton). He describes pressure from the fluid of the accessory glands. The gonapo- that in this species, the male first climbs on the female, physes of Tenthredinoidea thrust alternately, and sperm vibrates his wings for a few seconds and then moves off the must be transferred mostly by this motion since phallic struc- female, his body now at a right angle relative to that of the tures are absent.’ But since he neither gives an account of female and his hind legs on her thorax. He probes the meso- how he derived at this description nor cites any evidence, sternum with his genitalia and moves towards the end of her it seems that parts of it are assumptions drawn from the abdomen, his hindlegs remaining on top of the female. He morphology. then positions himself so that their heads face in opposite directions. While inserting his aedoeagus, his hindlegs are on Possible movements of the volsella her abdomen and wings. Gordh notes that the coitus lasted 70 s on average. After about 30 s of copulation, during which Muscle qr pulls the distivolsella towards the basivolsella. If the pairs remained motionless, several females started to the basivolsella were a simple sclerotic plate, the parossiculus move, dragging the males backwards. Some females tried to would assume the shape of a semi-circle (in lateral view). But dislodge the males by pressing their hindlegs against the male because it is strengthened through the carina volsellaris, the genitalia. basivolsella retains its shape, and the distivolsella is bent against the basivolsella, forming an angle of approximately 70° between them (Fig. 9C). The gonossiculus follows this Function of the male genitalia movement, since it is immovably fused along its middle part Enslin (1912) and Boulangé (1924) assumed that the with the distivolsella. Figure 9C shows the copulatory aedoeagus is introduced into the female genital opening at apparatus of a male of A. lichtwardti taken in copula. The the base of the saw. This and the assumption that the volsellae volsella on the right side of the drawing is in the relaxed grasp the edge of the female sternum (Boulangé 1924) is position, in which the volsella is slightly open, whereas the shown to be correct in the present study. The observed defor- volsella on the left shows the closed condition in which the mation of the female sternum shows that the harpes, too, distivolsella is pulled towards the basivolsella. (Note that it serve a grasping function, contrary to the belief of Boulangé can only be seen from a different angle whether the volsellae (1924; p. 100), who thought they would only be used to feel are open or closed.) When the right, relaxed distivolsella or for and maybe excite the female. muscle p (of the distivolsella) was pulled basally with a pin- Boulangé (1924) assumed that the gonomacula can act as cer, the digiceps moved towards the distivolsella so that the a suction cup, but was unable to prove this. He rejected volsella closed, even though the digitus had not been moved Crampton’s (1919) theory that the gonomacula was a ‘sen- directly. I do not yet understand the mechanism by which sory area’, because he found that the nerve leading into the this movement is achieved. harpe does not reach the gonomacula. Here it is demon- When the digiura of the relaxed volsella was pulled basally strated (in Macroxyela) that the gonomacula can indeed be – simulating the action of muscles s and si – the digiceps used as a suction cup. The fact that a pulsating action of the moved further away from the distivolsella, so that the volsella gonomaculae was observed is an indication that the attach- opened more than in the relaxed position. ment is indeed achieved through suction and not through a sticky substance secreted by glands. Boulangé (1924) Description of presumed muscular actions during the copulation in depicted thin sections of the gonomacula of Sirex juvencus strophandrous Tenthredinidae (Linnaeus 1758) which do not show any glands. D’Rozario (1940) observed in several pairs of Nematus When the male walks backwards towards the female, he ribesii (Scopoli 1763) fixed in copula that ‘the penis’ (mean- pushes his copulatory organ out of the genital chamber prob- ing aedoeagus) ‘is inserted deep into the vagina and its valves ably by haemolymphic pressure and the help of muscles c. open wide apart while its tip lies well under the spermathecal He opens the latimeres by contracting muscles d and f, opening into the vagina.… The sperm (s) are ejaculated in a reaches around the ventral side of the female abdomen, and mass into the vagina (v) along with the secretion (agf ) from – after finding her genital opening – contracts e and g to pull the male accessory gland (…)’. the latimeres together and t′′′ and t′′′′′′ to bend the harpes. At the E. L. Smith (1970a) wrote: ‘The male grasps the female same time or shortly after, the male moves the penisvalvae by from above or behind, depending on whether the genital cap- muscles h, i, j and k into the female genital orifice. He opens sule is rotated. The female genital orifice between the gona- the volsellae by contraction of s, possibly moves them around pophyses is exposed either voluntarily or through pressure by using o and p, and – upon sensing (presumably with the the male on her gonocoxites, and the volsellae appear to parossicular bristles) that he found the right place – closes clutch or hook the labia. The gonapophyses (and penis, if them around the seventh sternum by contracting muscles

Acta Zoologica (Stockholm) 82: 331–349 (October 2001) Schulmeister • Male genitalia and copulation in Hymenoptera

p and qr. He might then contract muscles n to open the men and genitalia in a complicated manner to reach the primary gonopore and pull apart the penisvalvae in the female female genital opening. Drawings, photographs and descrip- by contracting the muscles l. These actions could serve to tions of the copulation in three species of Xiphydria show that open the secondary gonopore (phallotrema) and to make the male clings to the very tip of the female abdomen, with way for the sperm. The male might push the sperm out by his abdomen upside down under hers, not giving the impres- alternately moving the penisvalvae by muscles h, i, j and k. sion of a secure hold on the female (Rohwer 1915; Chrystal At the end of the copulation, the gonostipites and harpes and Skinner 1932; Jänicke 1981). It can be assumed that the are opened through relaxing the muscles t′′′, t′′′′′′, e and g, and finding of the female genital opening could be rather difficult contraction of muscles u, d and f. The volsellae are opened when the male has to grope around with a twisted abdomen, to release the female seventh sternum by relaxing o, p and qr which could mean losing energy and time, the latter enhanc- and contraction of s. The copulatory organ is pulled back ing the probability of becoming a victim of predators during into the genital chamber by relaxation of c and the haemo- the copulation. E. L. Smith (1970b) mentions that some lymphic pressure and contraction of muscles a and b. Tenthredinoidea are carnivorous and indicates that strophandry would prevent the male from being attacked by the female, but since the ancestral species which evolved Ways of holding the female in a strophandrous species strophandry was probably not carnivorous, this cannot have During copulation of strophandrous as well as orthandrous played a role in the evolution of strophandry. species, the female often moves around (own observations in T. temula, E. koehleri and C. pygmeus). In a species in which Orthandry, strophandry and facultative strophandry the male is on top of the female during copulation this is not much of a problem, contrary to strophandrous or faculta- As mentioned above, the Xyelinae and Tenthredinoidea s. str. tively strophandrous species, in which the moving female are strophandrous, all other Hymenopterans orthandrous. drags the male behind her. In these species, a strong attach- In the present study, it was found that the orthandrous ment of the male to the female is therefore often essential for species Macroxyela ferruginea copulates in the manner of copulatory success. strophandrous sawflies, i.e. end-to-end. I will call this In Aglaostigma lichtwardti, a number of different measures facultative strophandry. Similar observations had already ensure this attachment no matter in which direction the been made by Eidt (1965) in the orthandrous Cephalcia female moves. (1) Pressing the harpes against the sides of the fascipennis (Cresson) (Pamphiliidae): ‘Without courtship, he female abdomen provides hold if the female moves side- mounts, then turns, all very quickly, so that the pair in copula ways. (2) Grasping the edge of the seventh sternum with the face in opposite directions in the manner of the strophandri- volsellae ensures attachment during forward movement of ous sawflies. This explains the specimen of A[cantholyda] the female. (3) insertion of the parapenes above the seventh burkei with the genitalia rotated 180° found by Middlekauff sternum prevents the female from pulling her abdomen (1958) who adds that it was taken in copula. However, upwards. (4) placing his hindlegs on her wings prevents her mated males of C. fascipennis and other species, when mating from pulling her abdomen upwards and from opening her terminated naturally, had their genitalia returned to the wings to fly away (which is probably supported through the upright position of orthandrious sawflies.’ and by Borden et al. plantulae of the hind tarsi). (5) The male’s position behind (1978) in Cephalcia lariciphila (Wachtl 1898) or C. alpina and under the female prevents backwards and downwards (Klug 1808): ‘... the female turned head-to-head with the movement. All in all, the male’s attachment is ensured male, then both turned or moved past each other and against movements in any possible direction. coupled tail-to-tail.’ This means that at least one macroxyelid and three pamphiliid species exhibit facultative strophandry, whereas Possible evolutionary advantage of strophandry members of the orthandrous groups Cephidae (own obser- In species copulating end-to-end, the holding of the female vation in Cephus pygmeus), Xiphydriidae (Rohwer 1915; has to be accomplished through additional measures. This Jänicke 1981), and Siricidae (Jänicke 1978) were observed leads us to question what the evolutionary advantage of copulating with the male clinging to the abdomen of the this copulation posture could be. In orthandrous and female. For the other orthandrous groups, namely the strophandrous copulation, the orientation of the male Blasticotomidae, Megalodontesidae, Anaxyelidae, and genitalia in relation to the female genitalia is the same; the Orussidae, the copulation posture is not known. anatomically ventral side (which carries the volsellae) always The fact that the male of C. fascipennis first mounts the faces upside during copulation. The orientation of the geni- female and then turns around is an indication that the male talia can therefore not have played a role in the evolution of might not be able to revolve his genitalia by himself. He prob- strophandry; the advantage must lie in the position of the ably attaches his genitalia firmly to the female (by using the male. My own observations of the orthandrous species gonomaculae) so that they become rotated automatically Cephus pygmeus showed that the male has to twist his abdo- while he turns around. Unfortunately, I cannot remember

Male genitalia and copulation in Hymenoptera • Schulmeister Acta Zoologica (Stockholm) 82: 331–349 (October 2001)

whether the males of Macroxyela mounted the female first Boulangé, H. 1924. Recherches sur l’appareil copulateur des before assuming the end-to-end posture, but I doubt that this Hyménoptères et spécialement des Chalastogastres. – Mémoires et was the case. The observation of Borden et al. (1978) indi- Travaux de la Faculté Catholique de Lille 28: 1–444. Chrystal, R. N. and Skinner, E. R. 1932. Studies in the biology of the cate that in some species the males might be able to actively woodwasp Xiphydria prolongata Geoffr. (dromedarius F.) and its turn their genitalia around. parasite Thalessa curvipes Grav. – Scottish Forestry Journal 46: 36–57. Even though Siricidae are orthandrous and Sirex juvencus Clausen, R. 1938. Untersuchungen über den männlichen Copula- (Linnaeus 1758) has been observed to copulate in the corres- tionsapparat der Ameisen, speziell der Formicinae. – Mitteilungen ponding manner (Jänicke 1978), Rasnitsyn (1969) notes der schweizerischen entomologischen Gesellschaft 17: 233–246. that a live Xeris spectrum male was capable of actively turning Crampton, G. C. 1919. The genitalia and terminal abdominal struc- its genitalia 180° and back. tures of males, and the terminal structures of the larvae of ‘chalas- togastrous’ Hymenoptera. – Proceedings of the Entomological Society We could now ask the question of how strophandry and of Washington 21: 129–151. facultative strophandry evolved. Unfortunately, if we code d’Rozario, A. M. 1940. On the mechanism of copulation in Nema- three different character states (O, FS, and OS), code the tus ribesii Scop. (Hym. – Proceedings of the Royal Entomological orthandrous taxa of which the copulation posture is Society of London. – Series a, General Entomology 15: 69–77. unknown as ‘O or FS’, and optimize this character on the Dufour, L. 1841. Recherches anatomiques et physiologiques sur les cladogram of lower Hymenoptera in Vilhelmsen (2001), Orthoptères, les Hyménoptères et les Névroptères. – Mémoires de this does not result in a single most-parsimonious scheme. If L‘Academie (Royale) Des Sciences de L‘Institut de France, Sav. Etrang. 7: 265–647. the Blasticotomidae were truely orthandrous, it would be Dufour, L. 1854. Recherches anatomiques sur les Hyménoptères more parsimonious to assume orthandry at the base of the de la famille des Urocérates. – Annales de la Science Naturelles et hymenopteran tree; if they were facultatively strophandrous, Zoologiques 1: 201–236. we would have to assume this state in the groundplan of the Eidt, D. C. 1965. The Life History of a Web-spinning Sawfly of Hymenoptera (both under the assumption that the out- Spruce, Cephalcia fascipennis (Cresson) (Hymenoptera: Pam- groups are orthandrous). To determine whether the ancestor philiidae). – Canadian Entomologist 97: 148–153. Enslin, E. 1912. Die Tenthredinoidea Mitteleuropas 1. – Deutsche of the Hymenoptera was orthandrous or facultatively entomologische Zeitschrift, Beiheft Suppl. 1. strophandrous, it would also be important to know what Gauld, I. and Bolton, B., eds 1988. The Hymenoptera. British copulation posture must be assumed at the base of the Museum of Natural History/Oxford University Press, Oxford. Mecopterida and the Coleoptera + Neuropterida. Gordh, G. 1975. Sexual behavior of Hemitaxonus dubitatus (Norton). – Entomological News 86: 161–166. Gibson, G. A. P. 1985. Some pro- and mesothoracic structures Acknowledgements important for phylogenetic analysis of Hymenoptera, with a I want to thank Rainer Willmann for suggesting the topic review of terms used for the structures. – Canadian Entomologist and supporting my work. I thank Malte Jänicke, Stefan 117: 1395–1443. Schmidt, David R. Smith and Lars Vilhelmsen for providing Goulet, H., Huber, J. T. 1993. Hymenoptera of the world: An identifi- me with valuable specimens. David R. Smith kindly helped to cation guide to families, Publication 1894/E. Centre for Land and Biological Resources Research, Ottawa, Ontario, Canada. – find and catch Macroxyela ferruginea. I am indebted to Lars Research Branch, Agriculture Canada, Ottawa. Vilhelmsen, Volker Mauss, Rainer Willmann, Jim Carpenter, Jänicke, M. 1978. Beitrag zur Biologie der Holzwespen (Siricidae) and Ward Wheeler for reviewing the manuscript and making I. – Veröffentlichungen Des Museums in Gera, Naturwissenschaftliche helpful suggestions. I especially thank Oliver Niehuis for his Reihe 6: 79–81. indispensable help and support. Jänicke, M. 1981. Beitrag zur Biologie der Holzwespen (Siricidae) II. – Veröffentlichungen Des Museums der Stadt Gera, Naturwissen- schaftliche Reihe 9: 79–82. References Kimsey, L. S. and Bohart, R. M. 1990. The Chrysidid Wasps of the Alam, S. M. 1952. Studies on ‘skeleto-muscular mechanism’ of the Wo r l d . Oxford University Press, Oxford. male genitalia in Stenobracon deesae Cam. – Beiträge Zur Entomo- Kluge, M. H. E. 1895. Das männliche Geschlechtsorgan von Vespa logie 2: 620–634. germanica. – Archiv für Naturgeschichte (a) 61: 159–198. Audouin, M. 1821. Observations sur les organes copulateurs mâles Königsmann, E. 1976. Das phylogenetische System der Hymenop- des Bourdons. Rapport lit par Latreille en lundi 9 avril. – Annales tera. – Teil 1: Einführung, Grundplanmerkmale, Schwestergruppe Générales Des Sciences Physiques 8: 285–289. und Fossilfunde. – Deutsche entomologische Zeitschrift 23: 253–279. Beck, D. E. 1933. A morphological study of the male genitalia of Königsmann, E. 1977. Das phylogenetische System der Hymenop- various genera of bees. – Proceedings of the Utah Academy of Sciences tera. – Teil 2: ‘Symphyta’. – Deutsche entomologische Zeitschrift 24: 10: 89–137. 1–40. Birket-Smith, S. J. R. 1981. The male genitalia of Hymenoptera – Kopelke, J.-P. 1982. Funktion der Genitalstrukturen bei Bombus- a review based on morphology in Dorylidae (Formicoidea). – Arten am Beispiel von B. lapidarius (Linneus 1758) und deren Entomologica Scandinavica Supplement 15: 377–397. Bedeutung für die Systematik. – Senckenbergiana Biologica 62(1981): Borden, J. H., Billany, D. J., Bradshaw, J. W. S., Edwards, M., Baker, R. 267–286. and Evans, D. A. 1978. Pheromone response and sexual beha- Michener, C. D. 1944a. A comparative study of the appendages of viour of Cephalcia lariciphila Wachtl (Hymenoptera: Pamphilii- the eighth and ninth abdominal segments of insects. – Annals of the dae). – Ecological Entomology 3: 13–23. Entomological Society of America 37: 336–351.

Acta Zoologica (Stockholm) 82: 331–349 (October 2001) Schulmeister • Male genitalia and copulation in Hymenoptera

Michener, C. D. 1944b. Comparative external morphology, phylog- genitalia. 2. Hymenoptera. – Annals of the Entomological Society eny, and a classification of the bees (Hymenoptera). – Bulletin of of America 63: 1–27. the American Museum of Natural History 82: 151–326. Smith, E. L. 1970b. Hymenoptera. In Tuxen, S. L. (Ed.): Taxono- Michener, C. D. 1956. Hymenoptera. In Tuxen, S. L. (Ed.): mist’s Glossary of Genitalia in Insects. 2nd enlarged Edition. Taxonomist’s Glossary of Genitalia in Insects. Munksgaard, Munksgaard, Copenhagen. Copenhagen. Smith, E. L. 1972. Biosystematics and morphology of Symphyta. 3. Middlekauff, W. W. 1958. The North American sawflies of the external genitalia of Euura (Hymenoptera: Tenthredinidae): genera Acantholyda. Cephalcia and Neurotoma (Hymenoptera, sclerites, sensilla, musculature, development and oviposition Pamphiliidae). – University of California Publications in Entomology behavior. – International Journal of Insect Morphology and Embryology 14(2): 51–173. 1: 321–365. Middlekauff, W. W. 1964. Notes and description of the previously Snodgrass, R. E. 1935. Principles of Insect Morphology. McGraw-Hill, unknown male of Syntexis libocedrii. – Pan-Pacific Entomologist New York and London. 40: 255–258. Snodgrass, R. E. 1941. The male genitalia of Hymenoptera. – Ohl, M. 1996. Die phylogenetischen Beziehungen der Sphecinae Smithsonian Miscellaneous Collections 99: 1–86. aufgrund morphologischer Merkmale des Exoskeletts. – Zoolo- Snodgrass, R. E. 1957. A revised interpretation of the external gische Beiträge (N. F.) 37: 3–40. reproductive organs of male insects. – Smithsonian Miscellaneous Peck, O. 1937. The male genitalia in the Hymenoptera (Insecta), Collections 135(6): 1–60. especially the family Ichneumonidae. – Canadian Journal of Thomson, C. G. 1871/72. – Hymenoptera Scandinaviae vol. 2, Research, Section D, Zoological Sciences 15: 221–274. p. 280. Ohlsson, Lund. Rasnitsyn, A. P. 1969. Origin and evolution of lower Hymenoptera. Togashi, I. 1970. The comparative morphology of the internal – Trudy paleontologicheskii Instituta Akadem. Nausk. SSSR 123: reproductive organs of the Symphyta (Hymenoptera). – Mushi 43 1–196. (Suppl.): 1–114. Rohwer, S. A. 1915. The mating habits of some sawflies. – Proceed- Verhoeff, C. 1893. Finden sich für die laminae basales der ings of the Entomological Society of Washington 17: 195–198. männlichen Coleopteren Homologa bei den Hymenopteren? – Ronquist, F. and Nordlander, G. 1989. Skeletal morphology of an Zoologischer Anzeiger 16: 407–412. archaic cynipoid, Ibalia rufipes (Hymenoptera: Ibaliidae). – Vilhelmsen, L. 1997. The phylogeny of lower Hymenoptera, with a Entomologica Scandinavica Supplement 33: 1–60. summary of the early evolutionary history of the order. – Journal Ronquist, F., Rasnitsyn, A. P., Roy, A., Eriksson, K. and Lindgren, of Zoological Systematics and Evolutionary Research 35: 49–70. M. 1999. Phylogeny of the Hymenoptera: a cladistic reanalysis of Vilhelmsen, L. 2001. Phylogeny and classification of the extant basal Rasnitsyn’s (1988) data. – Zoologica Scripta 28(1–2): 13–50. lineages of the Hymenoptera (Insecta). – Zoological Journal of the Ross, H. H. 1937. A generic classification of the nearctic sawflies. – Linnean Society 131: 393–442. Illinois Biological Monographs 15(2): 1–173. Willmann, R. 1981. Exoskelett der männlichen Genitalien der Ross, H. H. 1945. Sawfly genitalia: terminology and study tech- Mecoptera. – 1. Morphologie. – Zeitschrift für Zoologische Systematik niques. – Entomological News 56: 261–265. und Evolutionsforschung 19: 96–174. Schedl, W. 1991. Hymenoptera. – Unterordnung Symphyta: Pflan- Willmann, R. 1989. Evolution und phylogenetisches System der zenwespen. Handbuch der Zoologie 4(31). De Gruyter, Berlin, New Mecoptera (Insecta: Holometabola). – Abhandlungen der Sencken- York. bergischen Naturforschenden Gesellschaft 544: 1–153. Smith, D. R. 1990. A synopsis of the sawflies (Hymenoptera, Symphyta) Willmann, R. 1998. Advances and problems in insect phylogeny. – of America south of the United States: Pergidae. – Revta Bras. Ent. In: Fortey, Thomas, R. (Eds): Arthropod Relationships, pp. 269– 34: 7–200. 279. Chapman & Hall, London. Smith, E. L. 1969. Evolutionary morphology of external insect Wong, H. R. 1963. The external morphology of the adults and ulti- genitalia. 1. Origin and relationships to other appendages. – mate larval instar of the larch sawfly, Pristiphora erichsonii (Htg.) Annals of the Entomological Society of America 62: 1051–1079. (Hymenoptera: Tenthredinidae). – Canadian Entomologist 95: Smith, E. L. 1970a. Evolutionary morphology of external insect 897–921.