Male terminalia of Cercopidae (Hemiptera, Cicadomorpha): towards a consensus terminology

The study of male genital appendages is often necessary to identify a species and to characterise the higher systematics ranks for the Cercopidae, a large family of Hemiptera. Therefore, many authors have used them in their work but without any clear consensus on the terms used for each part constituting the male terminalia. A standardised terminology is important for the quality of a taxonomic description but even more essential when we want to compare species and establish a primary homology between states of character and their use in the frame of phylogenetic analysis. The use of a consensus terminology should ensure that we are all observing, speaking and describing the same genital appendage and comparing homologous characters. In order to propose a consensus terminology, we have reviewed all the major works on the anatomy of terminalia for the family since the first description using those characters in 1922. We proposed the use of consensual terms, listed with their definitions. In addition we studied a diversified panel of male specimens, chosen in order to represent as many Cercopidae tribes as possible. We categorised five different groups of Cercopidae according to their male terminalia structures. This opens the reflection on the evolutionary patterns for these structures.

In 1961 31 , the family Cercopidae was classically divided into two subfamilies: Cercopinae Leach, 1815 29 and Callitettixinae Metcalf 1961 31 . In 1968, Fennah 7 was working on Cercopidae from the New World and grouped them according to the structures of male and female genitalia. He suggested subdividing the family into two subfamilies, the Tomaspidinae containing all the New World genera and the Cercopinae dedicated to the Old World genera. He ignored the Callitettixinae because he was working on the New World taxa. According to Fennah, the Tomaspidinae generally present a metatibiae bearing two lateral spines, the male subgenital plates appear as an extension of the pygofer without any separation or groove and the first female valvulae bear some basal processes. In 2001, Hamilton 32 described a new family from the American tropics, the Epipygidae and suggested including the Aphrophoridae as a subfamily in the Cercopidae. Cryan and Svenson in 2010 33 presented the first phylogenetic molecular investigation on the Cercopoidea, which suggests that four of the five described families, Cercopidae, Clastopteridae, Machaerotidae and Epypigidae are monophyletic, the latest being nested within the Aphrophoridae. According to the selection of genera present in the phylogeny, a monophyletic lineage for the New World can be observed and designated as subfamily Ischnorhininae (sensu Carvalho 34 indicate a need to revise the genera included in the Tomaspidini and Ischnorhinini since those tribes do not remain monophyletic in their molecular analyses. According to Liang and Webb (2002) 3 , the current Old World classification of Cercopidae, together with many generic concepts, is based on the work of Lallemand 11 , using the number of spines on the hind tibia, and characters of the head, pronotum and fore wings. The Cercopinae is composed of numerous tribes and a series of incertae sedis genera. The monophyly of this subfamily was never tested. Liang and Webb 3 revised the Rhinaulacini from southern Asia. It is the latest morphological work done at a tribal level for the Cercopinae. Cryan and Svenson 33 point out that the taxonomic sample included in their phylogeny did not allow a comprehensive examination of Old World tribal structure, but some trends are emerging and suggest that their generic constituency needs to be examined in greater detail. If the construction of a molecular phylogeny for the Old World genera is needed in order to better understand their relationships, and the study of morphological structures should not be neglected. In order to be able to compare those structures we have to be sure that we compare the same structures. Male terminalia are no exception.
Results and discussion on the general morphology of male terminalia (Fig. 1). In 1910, Jacobi 9 valued the inclusion of the male terminalia description and their illustration, through the description of a series of new species of Cercopidae. This practice is now widespread among the authors and became a necessity in the description of each taxon. In the meantime, this generated numerous terminologies. For the Hemiptera different scientists made some attempt to homogenise the vocabulary used in this group, such as Crampton 1922 13 , Singh-Pruthi 1925 12 , Snodgrass 193514 , Kramer 195015 , and Marks 1951 . As presented in Table 1, no consensus has been reached yet. The absence of standardisation in the terminologies can lead to misidentification, confusion in identification keys or homology recognition, difficulty in the communication between scientists, and inefficiency in scientific results (Bourgoin et al. 2015) 35 . If for some terms, such as pygofer, subgenital plates and aedeagus a consensus seems to have emerged, others are still described under different names. Finally, other structures were mentioned recently since their presence was observed only in few taxa. This is the case for the lateral and intermediate plates, respectively mentioned by Liang and Webb 3 and Soulier-Perkins and Le Cesne 24 . We propose here a consensual terminology regarding the practices of authors in this domain (Table 1).
Terminalia. We group under this term the pygofer and the structures it bears plus the anal tube.
Pygofer (Fig. 1A). Here we consider it as being the ninth abdominal segment. However certain authors introduce some subtlety. Sing-Pruthi 12 refers to the pygofers as the large and conspicuous lateral regions of the ninth abdominal segment. Crampton 13 refers to the ninth sternite when he uses the term hypandrium. In the Cercopidae, it is a ring-like structure composed of the tergite and sternite, excluding the subgenital plates appearing as an appendage of it.
Anal tube (Fig. 1A). It corresponds to the tenth and eleventh segments. The sclerites of the tenth and eleventh segments are separated by inter-segmental membrane) and terminate in an elongated spoon-shaped process under the anus, the anal style as described by Singh-Pruthi 12 .
Subgenital plates (Fig. 1A). They are a pair of plates arising from the posterior ventral margin of the pygofer as referred to by Hamilton and Morales 36 . Sing-Pruthi 12 refers to them as a pair of well-developed appendages of the ninth sternite. They can be flexible but are never provided by muscles. In some groups, these appendages have developed in continuity with the pygofer and are fused to it. Some intermediate cases can be observed as well. The three alternatives can be observed for the Cercopidae.
Parameres (Fig. 1A,B). Crampton  www.nature.com/scientificreports/ genital styles as synonyms. Snodgrass calls them harpagones and describes them as a paired structure articulated to some part of the ninth segment. Individually provided with muscles, they arise from the floor of the genital chamber and are connected to the sclerites of the phallobase via the connectives. These appendages are generally placed on each side of the aedeagus.
Aedeagus (Fig. 1A,B,D). It is a median tubular structure 12 unequally chitinised. Some authors call the aedeagus plus the phallobase, the penis, or phallus. When the theca of the phallobase is developed and completely or D.
C. www.nature.com/scientificreports/ partially covers the aedeagus, they can be called the aedeagus sensus lato, the aedeaegus sensus stricto being only the aedeagus itself (Aed). In the case of the Cercopidae, the theca is completely reduced and the term aedeagus refers to the aedeagus sensus stricto. It communicates to the body cavity by the ejaculatory duct, which enters through the basal foramen 12 . This basal foramen is located in between the two basal sclerites, the basal plates, nested in the body wall or genital chamber. These sclerotic parts sometimes form a ring from which the aedeagus projects and provide attachments to phallic muscles. It is what we generally observe for the Cercopidae 14 . The whole aedeagus is sclerotized except for the most apical parts that seems to correspond to the endophallus. At the end, the gonopore opens.
Phallobase (Fig. 1A,B,D). As defined by Tuxen 4 , it is the whole structure supporting the aedeagus. For the Cercopidae it is reduced to a small ring at the base of the aedeagus. It is composed of the basal plates (BsP), nested in the genital chamber, extending toward the parameres and the thickened segmental membrane of the IXth segment covering them. Tuxen also mentions the connective as being a sclerotized structure belonging to the phallobase and connecting it to the styles and mentions that for the Auchenorrhyncha, it is a synonym for basal plate. It is Snodgrass in 1935 14 , who describes the phallobase for the first time as being the proximal part of the phallus highly variable in its development, sometimes a large structure supporting the aedeagus, often produced in a thecal fold of sheath about the aedeagus, sometimes represented only by basal phallic sclerites in the wall of the genital chamber. Since then, in descriptive works, the sclerified basal parts,extending toward the parameres took different names. Yang and Chang in 2000 37 , call it basal part phallobase, when Carvalho and Webb 6 call it connective.
Sterno-lateral plates (Fig. 1C). They are a pair of plates present only in some genera of Cercopidae. First described by Liang and Webb 3 , they seem to have the same origin as the subgenital plates, they arise from the ninth sternite and are generally flexible but are not provided with muscles. They are located above the subgenital plates, on the pygofer margin.
Intermediate plates (ItP, Fig. 1C). These structures are paired like the subgenital and lateral plates. They were observed first in 2016 by Soulier-Perkins and Le Cesne 24   www.nature.com/scientificreports/ only the tergite is considered the pygofer. The sternite, which may be fused or articulated to the tergite, is called the valve, and the subgenital plates, which also may be either fused or articulated to the valve (IXth sternite) are usually not considered part of the pygofer. In the Fulgoromorpha the term pygofer includes tergite and sternite 14,51 and there is no question about the subgenital plates that are absent in this group. In the Cercopoidea, because the tergite is fused to the sternite, it seems logical to call this ring like structure pygofer and consider the subgenital plates separately, like for the Membracoidea. Those subgenital plates are a pair of plates arising from the IXth sternite for all the cercopoids [38][39][40][41][42][43][44] , Cicadellidae 45,46 and Membracidae [47][48][49][50] . The term lateral plate is used for the Cercopidae for the first time by Liang and Webb 3 then Soulier-Perkins and Kunz 23 . These plates are present in some tribes of the old world Cercopidae, they arise like the subgenital plates from the IXth sternite. However, the term lateral plate is also used for the Membracidae 47-50 and designates a synapomorphic character for this group and this distinctly delimited structure, either entirely or distally free or entirely fused to the pygofer, originating from the IXth tergite. So, if in both groups we consider that the lateral plates arise from the pygofer, they are not of the same origin, and are not of homologous structures. For these reasons, we suggest changing the term lateral plate to sterno-lateral plate for the Cercopidae in order to emphasise their different origin. The appendages on each side of the aedeagus are generally named parameres in Cercopidae 6,12,15,22,23 and Machaerotidae 52 when the term style is preferred for the Aphrophoridae 39-43 , Clastopteridae 42,44 , Epipygidae 38 and Membracoidea [46][47][48][49][50] . Even though the term style is used for the Fulgoromorpha the term gonostyle is widely used as well 51,53 . Authors generally describe the aedeagus as the median tubular part bearing the ejaculatory duct. To refer to the sclerotized part the term theca is sometimes used, for instance for the Clastopteridae by Hamilton 42 or aedeagal shaft for the Cercopidae 19,22 . At the base of the aedeagus, the reduced phallobase is essentially represented by a thickened membrane of the IXth segment and the basal plates, which support the ejaculatory duct and connect the phallic structure to the parameres. This structure is not always drawn and named in taxonomic works, but when it is represented it can be called connective 6,50 or just labelled as phallobase 32,37,42 . For some Fulgoromorpha such as the Tropiduchidae, Bourgoin and Huang 51 describe the phallobase sensus lato as being a synonym of the periandrium and what they describe as phallobase sensus stricto (sensus Fennah 1945 54 ) is a development of this structure folding around the aedeagus in an external sclerified phalotheca and an internal membranous endotheca. This phallobase sensus stricto can be considered absent in the Cercopidae. They described what remains as a thickening of the diaphragm which corresponds to what we called thickened segmental membrane of the IXth segment. Bourgoin and Huang also describe precisely what they call the connective sensus lato an internal composite structure. The first part is composed at the base by a small pregenital invagination of the diaphragm called the ventral support, followed by the body of the connective, which runs in the general cavity and finishes in a gutter shape structure. This first part is the connective sensus stricto and it is topped by the second part, the tectiform structure. In between these two parts the ejaculatory duct goes through and rest in the gutter-shaped part. The tectiform structure does not exist in the Cercopoidea and the fulgoromorphan' connective is difficult to synonymise with the term connective used for the Cercopoidae. In both cases the described structures link the parameres to the base of the phallus and support the ejaculatory duct but are not quite of the same origin. Focusing on the Cercopidae, and according to the definitions provided above and the observations done on specimens selected to illustrate the cercopid diversity, we observe different degree of fusion of the subgenital plates to the pygofer, the presence or absence of sterno-lateral plates and when those plates are present another pair of plates can sometimes be observed as well, the intermediate plates.
Fennah in his work of 1968 7 pointed out the importance of the genital characters to study the Cercopidae and considered them as not being influenced by parallel evolution. Without discussing the higher rank of the family, Fennah is the first to mention the different configuration of the male terminalia and that it could reflect the phylogenetic relationship of family. He expressed it by dividing the Cercopidae in two subfamilies but did not deepen it. His hypothesis was recovered by a first phylogeny of the Cercopidae of the world 33 , where the new world cercopids derived from the Old Word. They suggested that the genital characters reflect the same evolution. As such, a complete fusion of the subgenital plates with the pygofer is a derived state compared to a partially fused or distinctly separated subgenital plates to the pygofer. The presence of sterno-lateral plates is only observed in some genera found in the Old world and could be a derived state and a synapomorphy within a monophyletic lineage. The appearance of intermediate plates linking the sterno-lateral plate to the subgenital plate could be a step further within the evolution of this character, but those hypotheses remain mere speculation for now since the phylogeny of the whole Cercopidae family is not yet available. The optimisation of these patterns should be done when it becomes possible to test them. It has to be pointed at as well that when the subgenital plates are partially or completely fused to the pygofer, no sterno-lateral plates can be observed.
For now, as a practical tool, but without any intention to describe different states of evolution, we propose to group the male terminalia of Cercopidae in five categories ( Table 2). As illustration for the group 1, we can take the type-species, Cercopis sanguinolenta for which the subgenital plates are clearly distinct from the pygofer and do not possess any other plates, neither a sterno-lateral nor an intermediate plate ( Fig. 2A,B). The group 2 includes all the New World cercopids observed. It is characterised by the fusion between the subgenital plates and the pygofer, the subgenital plates appear as an elongation of the sternite IX without any distinction from it (Fig. 2C,D). Those terminalia do not possess any additional plates. The group 3 possess subgenital plates partly fused to the pygofer, however a fold is visible, on each side, a groove is running down but does not reach the ventral part (Fig. 2E,F). If the pygofer is generally shaped as a uniform sclerotized ring, in this third group, a ventral narrow antero-posterior membranous band can be observed. The group 4 presents some subgenital plates clearly distinct from the pygofer, some extra structures can be observed, above the subgenital plates. Those structures are the sterno-lateral plates (Fig. 2G,H). A groove individualises them from the pygofer. The group 5, similar to group 4, possess a pair of subgenital plates distinct from the pygofer, a pair of sterno-lateral plates and a pair of intermediate plates that are, small sclerotised plates, shaped as a bridge and linking the subgenital plate to the sterno-lateral plate (Fig. 2I,J).  www.nature.com/scientificreports/