Immature stages and ecology of two species of the South African genus Stripsipher Gory & Percheron, 1833 (Coleoptera, Scarabaeidae, Cetoniinae, Trichiini)

Abstract Based on the study of newly accessible type material, Stripsipher drakensbergi Ricchiardi, 1998, is demoted to a junior synonym of Stripsipher jansoni Péringuey, 1908. The genus Stripsipher Gory & Percheron, 1833, thus, currently includes 12 species, but for none of these are larval stages and/or pupae currently known. The immature stages of Stripsipher orientalis Ricchiardi, 2008 and Stripsipher jansoni are described here for the first time and updated observations on distribution and ecology of both species are provided. Morphological affinities of Stripsipher with other Trichiini larvae are presented and the main diagnostic differences discussed. The larvae of both species are very similar to those of other representatives of the tribe Trichiini, with key differences found on the epipharynx. Based on the morphology of larvae and adults, it is suggested that Stripsipher is a member of the clade composed of Valgini, Trichiini and Cryptodontini.


Introduction
Following the latest revision, the genus Stripsipher currently includes 13 species (Ricchiardi et al. 2008) (see list below). For one of these, S. jansoni Péringuey, 1908, it had not been possible in the past to examine any type specimen and, therefore, the position of incertae sedis had been preferred for the species. Recently, the lectotype (here desig-nated) has become accessible for analysis, courtesy of J Krikken of the Museum Naturalis, Leiden. Péringuey (1908) described this species from a pair, however the actual location of the female is unknown. This has now revealed that S. jansoni is actually a senior synonym of S. drakensbergi Ricchiardi, 1998. Therefore the current number of recognized species for the genus is reduced to 12.
An important contribution to this analysis could come from studies of larval characteristics of the species included in the genus. Unfortunately, there are still too few descriptions of Trichiini larvae and none currently known for the genus Stripsipher. Here, we provide the first description of the larvae of two species of this genus, which were conclusively identified through the rearing of cohorts collected in their natural environments.

Material and methods
The classification of the Cetoniidae used here follows Krikken (1984). A number of divergent classification systems have been proposed recently (e.g. Scholtz and Grebennikov 2005;Smith in Bouchard et al. 2011), but they remain equally controversial and not universally accepted.
The terminology for larval morphology follows Hayes (1929), Böving (1936), Ritcher (1966) and Sawada (1991). Hair-like setae of cranium and other structures were classified according to their relative size into two groups, medium to long (100-500 μm) and minute (20-50 μm, or less) setae, in order to give the most accurate information on chaetotaxy. Refer to Šípek et al. (2008;Fig. 12) for a detailed schematic figure. Morphological analysis and measurements were carried out using Olympus SZX9 and Olympus BX 40 light microscopes, both equipped with digital camera Olympus Camedia 5060. Mouthparts were dissected and if necessary mounted on slides in Liquide de Swan (e.g. Švácha and Danilevsky 1986). Photographs of beetles and larvae were taken using a Canon 550D digital camera equipped with a Canon MP-E 65/2,8 MACRO lens with 5:1 optical magnification. Final images were composed from multiple layer-focus pictures using Helicon Focus software. Structures examined at the scanning electron microscope JEOL 6380 were cleaned in 10% lactic acid for 24 hours, dried with critical point drying and mounted on aluminium plates. Drawings were made on the basis of photographs and all photographs were digitally enhanced (levels adjustment, background elimination, sharpening) using Adobe Photoshop.

Larval rearing
Eleven and 22 larvae of S. jansoni and S. orientalis were collected at Cobham (KZN southern Drakensberg) or Karkloof and Entumeni (KZN Midlands), respectively. Three to five specimens of each population were immediately fixed in 10% formalin, while the others were reared to adulthood in the laboratory, using their natural food sources (soil detritus and decomposing wood). They were kept in an environmental control room at the University of KwaZulu-Natal at a constant temperature of 22 o C during the summer, which was then gradually decreased during the colder seasons to a minimum of 15 o C in winter (July-September). Larvae were kept in 10-litre buckets filled to approximately half capacity with their natural soil/wood. The surface layer of each bucket was sprayed with tap water at regular weekly intervals until pupation of the bulk of the larvae, but left to dry out thereafter. The success rate was about 80% for both species, with adults emerging in good conditions over a period of about two months, from late October to December.

Phylogenetic analysis
To test the phylogenetic position of Stripsipher, a previously published larval/adult morphological dataset (Micó et al. 2008(Micó et al. , Šípek et al. 2009) was used, which was supplemented with new data. Methods of phylogenetic reconstructions follow Šípek et al. 2009. All bioinformatic analyses were carried out on the freely available Bioportal (www.bioportal.uio.no).

Stripsipher orientalis
This is a forest dwelling species (Figs 1A, C, E), which completes its larval development inside old tree trunks and branches, feeding on decaying, soft wood. It is distributed from the Eastern Cape north-eastwards, including KwaZulu-Natal, Swaziland, Mpumalanga, Gauteng, North-West and Limpopo (Fig. 2). Larvae have been found in a variety of tree species (e.g. Podocarpus spp., S. Endrödy-Younga and P. Istvan, specimen data labels), mostly in relict afromontane and coastal rainforests (e.g. Fort Fordyce, Umthamvuna, Karkloof, Entumeni, Monk's Cowl). Invariably, they attack old wood that has already been partly consumed by other saproxylic species, but is still suspended in the air or at least not directly in touch with the ground, presumably in order to avoid soil predators (e.g. ants). They appear to avoid entirely dying or recently dead trunks and branches. Larvae have been reared successfully in captivity, in environmental control rooms in Durban, by one of the authors (R.P.).
Adults of the species are active from September to February, exhibiting a peak of records during the months of December (33% of total) and November (26%). They have been observed flying in dense forests around decomposing wood during the hottest part of the day, or alternatively crawling on dead trunks/ branches and even in tunnels inside the wood. They feed on nectar, having been recorded on a variety of flower types, including Buddlia saligna, Sizigium cordatum, Dalbergia obovata, D. armata and Protea caffra. They have also been captured with fruit-baited traps on at least two occasions (S. Endrödy & M. Klimaszew and O. Bourquin, specimen data labels) and in Malaise traps (R. Miller, specimen data label), but not at sap flows.  2). An exception to this, possibly due to an error in identification, appears to be the record from the Mbotyi Forest (G. van Son, specimen data label) which is a coastal locality of elevation not exceeding approximately 500 m Larvae of this mountain dweller have been found under stones and in crevices among rocks, where they feed on accumulations of decomposing grass litter or herbivore dung (Ricchiardi et al. 2008). Particularly well utilized appear to be the pellets of Rock hyrax, Procavia capensis, locally known as "dassie", which at these high altitudes attract not only S. jansoni, but also scores of larvae of other Cetoniinae species (e.g. Xeloma spp., Rhinocoeta spp., Trichostetha spp. Rhabdotis spp., Leucocelis spp. Parelaphinis moesta, Hypselogenia geotrupina) Marais 1992, Holm andPerissinotto 2004, R.P. pers. observ.). Larvae from the Drakensberg (Cobham Nature Reserve) and the Kwamandlangampisi Mountain have been reared successfully in captivity, in environmental control rooms in Durban by R.P.
Reflecting its high altitude habitat, S. jansoni exhibits a delay in its adult activity, compared to S. orientalis. Records range from October to March, with peaks in December (48% of total) and January (24%). S. jansoni belongs to a species group within the genus that does not exhibit any feeding at the adult stage. Adults have been found crawling on the ground, flying low above the ground (10-50 cm) or resting on a variety of grasses. On one occasion, a female specimen was collected on yellow flowers of an unidentified shrub species on the Compassberg (R.P. pers. observ.), but it could not be established whether this was for the purpose of feeding or simply resting.  Larvae C-shaped, grub like, abdomen 10-segmented, segments of thorax and abdomen almost equal, abdominal segments 7-9 somewhat elongated. Length of larvae studied (third instars) 33-51 mm. Head capsule (Figs 3C, D): maximum width between 2.8 and 4.2 mm. Surface of cranium with faint microsculpture, orange brown to brown, antennifer, postclypeus and labrum brown. Cranial chaetotaxy summarized in Table 1. Frontal sutures straight, slightly crooked in the anterior part. Epicranial insertions of antennal muscles distinct, more sclerotised than the surrounding area of epicranium or only indicated as a small depression near middle of frontal sutures. Clypeus trapezoidal, anteclypeus narrow and membranous, postclypeus heavily sclerotised with one exterior and one anterior clypeal setae on each side. Frontoclypeal suture distinct. A single pigmented stemma present on each side.
Antennae (Figs 3C,D): antennomeres (an) subequal in length (relative length: an I = an IV > an II > an III; an II about 2/3 of an I and an IV). Third antennomere (an III) with ventral, apical projection exhibiting a single sensory spot. Last antennomere (an IV) with two or three ventral and one dorsal sensory spot and a small apical sensory field with faint sensillae.
Labrum symmetrical, anterior margin convex with numerous setae. Medial part with transverse, emarginated protuberance. Dorsal labral surface with two transverse rows of setae. Posterior row with about 10 setae, anterior row with one paramedian and one lateral seta on each side.
Epipharynx (Figs 3E-F, 4A-B): Haptomerum with strong haptomeral process and two to four transversal arched rows of stout setae. Haptomeral process prominent, well sclerotised with apical field of five pores and several others scattered around. Acroparia with about 30 long setae, clithra and epizygum absent. Lateral margin of epipharynx heavily sclerotised, especially in distal part. Plegmata and proplegmata absent. Acanthoparia with seven to eight setae. Chaetoparia asymmetric, with right exhibiting seven to eight irregular rows of setae and left five to six. Setae of medial rows stout, spine-like. Dexiotorma bar-like, straight, longer than laeotorma, right pternotorma almost or entirely absent. Laeotorma triangular with posterior pternotorma. Sense cone of haptolachus with four pores, sclerotised plate absent. Crepis faintly sclerotised, indistinct.
Mandibles ( Maxilla (Figs 5, 6A, B): dorsal surface of cardo and dorsal and ventral surfaces of labacobaria with hair-like setae and pores; vertral side of cardo bald. Dorsal surface of stipes with 10-20 slender hair-like setae in one to three longitudinal rows; oblique row of two to six well sclerotised spine-like stridulatory teeth and an anterior truncate process (blunt tubercle) distad to the row (Figs 5A, B, G, H, 6B). Stipes with few ventral setae. Galea and lacina entirely fused forming mala, galeo-lacinial suture dorsally visible as desclerotised line (Fig. 6B), entirely absent on ventral face (Fig. 6A). Galean portion of mala with single falcate uncus and several long and stout hair-like setae in longitudinal rows; lacinial part of mala with two reduced unci fused at their base and two apical stout spine-like setae (Figs 5E-F); dorsomedial side with numerous and very long hair-like setae. Ventral surface of mala with row of three to five stout setae and few long setae. Maxillary palps four-jointed, with basal joint often partially retracted into the palpifer, making palps appear three-jointed (Figs 5A, B, 6A, B). Penultimate joint usually with two setae.
Hypopharyngeal sclerome (Figs 5C, D) asymmetrical with strong protruding and pointed truncate process. Tufts of tegumentary expansions (= phoba, sensu Böving, 1936) present on left lateral lobe, hypopharyngeal sclerome proximal to truncate process and below hypofaryngeal sclerome, but present or absent near its right proximal border. Right lateral lobe not sclerotised.
Ligula (Figs 5C,D) dorsally with numerous setae, proximal setae stout conical and short, posterior setae hair-like, long. Dorsal surface of ligula with pit-like organ in Table 1    the medial portion and several pores in the distal half. Posterior margin with pair of long paramedian setae. Labial palpi two-segmented. Thorax (Figs 3A, B): Prothorax with single dorsal lobe, meso-and methatorax with three well developed lobes. Each dorsal sublobe of thoracic segments with one to three rows of short setae, posterior row with short setae interspersed with long setae. Prothoracic sclerite covering almost the whole lateral portion of prothorax, posterior border with median projection extending almost behind the distal margin of mesothoracic spiracle. Mesothoracic spiracle (Figs 6E, F) with C-shaped respiratory plate; distance between lobes of respiratory plate almost equal to the maximum diameter of respiratory plate. Respiratory plate with 12-25 holes across diameter. All pairs of legs (Figs 3A, B, 6C, D) subequal and with similar tarsal claws; claws falcate, sharp pointed and bearing two proximal setae (Fig. 6D).
Abdomen (Figs 3A, B, 6G, H): ten-segmented. Dorsa of abdominal segments I-VI with three sublobes, segments VII and VIII with only two. Each sublobe bearing three to five rows of short setae, posterior one or two rows with long setae interspersed between the short ones. Abdominal spiracles similar to mesothoracic spiracles, but slightly smaller. Size of spiracles decreasing caudad. Smallest spiracle on abdominal segment VIII, reaching about two thirds of the size of the mesothoracic one. Raster either without palidium and with fused teges composed of numerous hamate setae (septula absent, Fig. 6G), or with palidium consisting of paired oblique rows of pali (Fig. 6H). Each row of approximately ten pali, extending forward and inwards, thus forming a "V" or "U" on the medial portion of last abdominal segment, enclosing a small septula. Teges composed of few hamate setae. Ventral anal lip with 20-40 hamate setae, dorsal lip with several long hair-like setae; same setae present also on ventrolateral and ventroanterior portions of last abdominal segment.
Diagnostic characters of S. jansoni Péringuey, 1908 (Figs 3B, D, F;4B, D, F;5B, D, F, H;6B, F, H) (Figs 5B, 6B): stipes dorsal with one or two rows of approximately 10-18 slender hair-like setae and two to six conical stridulatory teeth (Fig. 5H). Ventral face of mala with row of three stout conical setae (occasionally another stout seta preceding the row, reaching about ½ of the size of the following three), lacinial unci obtuse. Ligula dorsal with 50-60 short conical to long hair-like setae (Fig. 5D). Size of thoracic spiracle 0.35-0.40 × 0.24-0.26 mm (length x width). Respiratory plate with 12-16 holes across diameter. Body vestiture more or less dense, setae on dorsum of ninth abdominal segment not separated by bald area into two fields. Palidium of raster consisting of paired oblique rows (Fig. 6H). Each row of ap-proximately ten pali, extending forward and inwards thus forming an inverted "V" or "U" shaped row on the medial portion of last abdominal segment, enclosing a small septula. Teges composed of few hamate setae. Two distinct types of hamate setae present on teges, small and large with broadened base. Large hamate setae about three times longer than pali.
Larvae of Osmoderma and Platygenia have fused abdominal segments IX and X, which is in contrast to what is found in all other known Trichiinae, which have a ten-segmented abdomen. Moreover, larvae of Osmoderma lack visible stemmata and the shape of their tarsal claw also differs significantly from that of all other Trichiinae.
The morphology of the epipharynx is amongst the most critical characters separating Stripsipher from all other non-African Trichiini (e.g. Medvedev 1952, Ritcher 1966, Morón 1983, Zhang 1984, Delgado-Castillo and Morón 1991, Sawada 1991, Morón 1995, Klausnitzer and Krell 1996. In Stripsipher (as well as in Clastocnemis and Coelocorynus) the haptomeral process is located distad to a transverse row of stout setae and the proximal end of dexiotorma is straight (Figs 7A, B). In all the remaining Trichiinae, on the other hand, the row of haptomeral setae is interrupted by a strong process and the dexiotorma is bent mesally at its proximal end (Fig. 7C).

African Trichiinae
As stated above, larvae of only four genera of Trichiinae are known from sub-Saharan Africa. Apart from body size, Platygenia larvae can also be separated from the rest by the fusion of abdominal segments IX and X. Larvae of Coelocorynus are characterized by strong lateral tubercles on epipharynx (Fig. 7B) and a drastic reduction of lacinial unci. The larvae of Clastocmenis quadrimaculatus seem to be the most similar to the larvae of Stripsipher. However, the description provided by Jerath and Unny (1965) is incomplete, so the only difference that can be derived from this lies in the number of lacinial unci. While in Clastocnemis there is a single lacinial uncus, two small unci can be observed in Stripsipher larvae.

Phylogenetic relationships
To evaluate the phylogenetic relationships of the genus Stripsipher, both larval and adult morphological characters of the two species were coded into a data matrix used in previous phylogenetic studies of the Cetoniinae (Micó et al. 2008(Micó et al. , Šípek et al. 2009). Not surprisingly, results do not differ significantly from those obtained in previous works. However, the genus Stripsipher proved to be firmly nested within the Trichiini-Cryptodontini-Valgini clade (Fig. 8). This finding supports the empiric observation  Micó et al. (2008) and Šípek et al. (2009). that the larvae of Stripsipher strongly resemble the known larvae of all other Trichiinae, with the exception of those belonging to the genera Osmoderma, Inca, Archedinus and, so far as can be derived from the brief description in Jerath and Unny (1965), also of the genus Platygenia. Regardless of the controversial position of the genus Valgus within a clade comprised of Trichiini and Cryptodontini members, this study supports once again the hypothesis that the Trichiinae (sensu Krikken, 1984) represent a paraphyletic group (see also Browne and Scholtz 1998;Micó et al. 2008;Smith et al. 2006).