Life history and description of larva and pupa of Platyphileurus felscheanus Ohaus, 1910, a scarabaeid feeding on bromeliad tissues in Brazil, to be excluded from Phileurini (Coleoptera, Scarabaeidae, Dynastinae)

Abstract The third instar larvae and the pupae of Platyphileurus felscheanus Ohaus, 1910 (Phileurini), recently synonymized with Surutu jelineki Endrődi, 1975 (Cyclocephalini), are described and illustrated, and some life history information is given. The larvae were collected and reared in bromeliads in rain forests of Santa Catarina state in southern Brazil. The systematic position of this monotypic genus is reassessed at the tribe level by considering larval and adult morphological characters. Both character sets, being described and illustrated, suggest the placement of Platyphileurus in the tribe Oryctini.


The enigmatic Platyphileurus felscheanus / Surutu jelineki
The rhinoceros beetle Platyphileurus felscheanus Ohaus, 1910 (Scarabaeidae: Dynastinae) was described twice as a new species, first under this name in the tribe Phileurini and 65 years later as Surutu jelineki by Endrődi (1975) in the tribe Cyclocephalini. Grossi et al. (2010) recently synonymized those two names.
The monotypic genus Platyphileurus Ohaus, 1910 is known only from Brazil (Grossi et al. 2010) with one erroneous record from the Ilha do Principe (West Africa) (Endrődi 1977). Grossi et al. (2010) extended the known distribution of the species from Santa Catarina (Ohaus 1910) and Rio de Janeiro (Endrődi 1985) to the states of Bahia, Espírito Santo, Minas Gerais and Paraná.
Platyphileurus felscheanus can be recognized by its flat body, especially anteriorly, by the elytra being laterally dilated posteriorly, by lacking horns or tubercles on the head and pronotum, and by lacking a longitudinal furrow on the pronotum . With the anterior half of the body flatter than the posterior half and the flattened pronotum lacking a longitudinal furrow, Platyphileurus has an unusual appearance for a phileurine species. Its systematic position needs to be re-examined.
Here we describe the larva and pupa of P. felscheanus repeatedly collected in bromeliad rosettes and reared to imagines, present data on their life history, and give new records on the occurrence of this bromelicolous species. We also explore whether characters of the immatures provide indications of the tribal placement of the genus.

Collection and rearing of beetles
All large beetle larvae and pupae found in the bromeliads were identified as scarabaeids. We prepared small rosettes of N. innocentii from the innermost part of the plant (about 8-10 young leaves), washed with tap water to remove spiders and other predatory arthropods, or arranged freshly cut, clean bromeliad leaves to an artificial rosette in a funnel. Larvae and pupae were placed in the middle of the rosette and covered with 1-2 table spoons of humic leaf litter. The arrangement was covered with gauze and kept moist in the laboratory. Larvae were inspected about every second week to check their vitality and to replace the eaten up rosettes or leaves with new ones. The identification of P. felscheanus was based on imagines obtained from these rearings.
In order to collect more specimens of P. felscheanus and to learn about its life history, five additional bromeliads of the genera Aechmea Ruiz & Pav., Nidularium Lem. and Vriesea Beer. were collected between 2008 and 2012 in Florianópolis. In the field, we searched bromeliads for scarabaeid larvae. When a larva was present (n=6), the bromeliad was taken to the laboratory where it was kept upright in a plastic bucket. During the first days or weeks after collection the larvae were left in the bromeliads to observe their behavior. Thereafter, some bromeliad leaves were tied together in small rosettes and each larva was placed in the middle of this artificial rosette which were maintained in plastic pots (n=5). One larva was maintained in the original bromeliad until pupation. They were checked one to three times per week.

Phylogenetic reasoning
A cladistics analysis at tribal or generic level is beyond the scope of this paper. In full consideration that the current tribal classification rests on entirely typological foundations, we apply consistently phylogenetic reasoning sensu Hennig (1982) and Watrous and Wheeler (1981) to interpret character states and their distribution to determine a possible placement of Platyphileurus in a current tribe. Ohaus, 1910 The larval description is based on four third instar larvae with the following data: BRAZIL, Santa Catarina: UCAD, Florianópolis city, in Canistrum lindenii, 18.ii.2002 Additional material: Two last larval instar exuvia and one pupal exuvia (MZSP) of the reared P. felscheanus larvae.

Platyphileurus felscheanus
To assess the differences of apical setal patterns of pupae between different species of Dynastinae, pupae of the following species of Phileurini and Oryctini from the immature collection of MZSP were studied (for descriptions see Vanin et al. (1983), Costa et al. (1988)): (
Left mandible (Figs 5,8): Form falcate. Scissorial area with S 1 and S 2 distant but bridged by flat area forming broad apical blade, separated from S 3 by acute scissorial notch. S 4 of similar size as S 3 , blunt, of cylindrical appearance in ventral view, separated from S 3 by acute and deep notch. Mandible dorsally with 1 long discal seta in front of labium at level of S 3 (arrow). Outer margin convex. Scrobis (SCR) with 1 short, thin seta. Dorsal area adjacent to scrobis with 2 rows of 7 sensorial pits. Dorsal area adjacent to molar crown with 3 setae. Acia (AC) well developed, with brush of apical setae. Brustia (BR) with 12 long setae. Ventral surface with elongate-oval stridulatory area (STA) with about 30 narrowly separated, subparallel ridges. Molar area with a tuft of 8 ventral molar setae (VMS) (setae very close together and difficult to count). Molar lobe (ML) large, forming a dorsoventral ridge, not subdivided. Molar crown with 2 lobes. Postartis (PTA) large, spherical. Ventral process triangular with rounded tip.
Right mandible (Figs 6-7): Form falcate. Scissorial area with S 2 separated from S 1 by obtusely angled notch. S 2 and S 3 distant but bridged by flat area, S 3 hardly developed as a denticle but with deep, acute notch separating it from S 4 . S 4 triangular, as elevated as S 2 . Mandible with 1 long discal seta in front of labium at level of S 3 (might be lacking or broken off) (Fig. 6, arrow). Outer margin convex. Scrobis with 1 short, thin seta (SCRS). Dorsal area of scrobis with an inner row of 5 and an outer row of 9 sensorial pits. Dorsal area adjacent to molar crown with a row of 4 white pits with 0, 4, 1, and 2 setae, respectively; discally with longitudinal row of 3 distinct, white pits with 3, 1, and 2 setae, respectively. Ventral surface with elongate oval, anteriorly tapering, stridulatory area (STA) with about 30, narrowly separated, subparallel ridges. Molar area with 5 ventral molar setae. Molar crown with 3 blunt ridges. Calx (CA) ventrally and dorsally ending in slightly blunt denticle. Brustia (BR) with about 15 setae. Postartis (PTA) large, spherical. Ventral process (VP) suboval, elongated laterally.

Maxilla and labium, dorsal view
Antenna (Fig. 4): Four-segmented with fourth antennomere the longest, about 1.7 times as long as third, second antennomere slightly shorter than fourth, first antennomere slightly shorter than third but thicker than others. Terminal antennomere with 12-13 ventral sensory spots (VSS) and 13-15 dorsal sensory spots (DSS); apex with 1 sensory spot.
Legs (Figs 11-12): Tarsal claws falcate, all similarly curved and similar in size, with 1 basal seta and 1 seta in the middle of inner side. Tibiotarsus (TT) with 4 apical setae on outer side of the base of claw, and with 2 circular rows of 6 long setae. Femur (FE) with 2 circular rows of 4-5 long setae and accessory setae. Trochanter (TR) with 5-6 ventral setae. Coxae (CX) with 4 setae. Setae of legs light brown to transparent, thin.
Abdomen (Fig. 11): Spiracles of similar size as prothoracic spiracle; last one smaller; rounder than prothoracic spiracle. One long seta on stigma area behind each abdominal spiracle II to VII, no seta behind spiracles I and VIII. Pleural lobes (PLL) with 2 long setae. Pedal area with 1 central and 2 lateral long setae per segment. Abdominal tergites (AT) I-VII with many tiny, short, dark, spike-like setae, not arranged in rows. Abdominal tergite IX laterally with dark, sparse minute, spike-like setae. Dorsum X (DO X) completely covered with such setae.

Description of the pupa of Platyphileurus felscheanus
Length 23.9 mm; largest width 10.4 mm. Adecticous; exarate; body oblong, smooth, apparently glabrous but with microsetae covering whole body (best seen at magnification > 50 ×), apex of tergite IX with dense tuft of setae seen in dorsal and ventral views; abdominal segments constitute almost two thirds of whole body; yellowish-brown before and after fixation, gin-traps and spiracular rings darker and more strongly sclerotized.

Life history
In 14 ground-growing bromeliads, we collected 19 larvae of P. felscheanus, 13 of them in Aechmea lindenii, three in A. nudicaulis, one in Canistrum lindenii, one in Hohenbergia augusta, and one in Nidularium innocentii. The highest number of larvae in one bromeliad was four in a single A. lindenii, but the majority of bromeliads had only one larva.
Second instars were found at the base of more external leaves of the rosette among leaf litter. No damage to bromeliad leaves near the larvae was observed. Third instars were always found in the center of the rosette with the surrounding leaves strongly damaged (Figs 32-33).
Coincidentally, when the bromeliad leaves started to dry and consequently to decompose, the larvae stopped feeding and started pupation. They built their pupal chamber in the center of the leaf rosettes or among leaves near the center, using leaf litter, twigs, and humus that had accumulated there. The larval exuvia was pushed to the rear of the pupal chamber (Fig. 34). Observations of the second phase are summarized in Table 1.
Parasitism by Tachinidae flies was observed in one pupa (Fig. 35). As the bromeliad with the larva was kept open in the laboratory, we do not know whether parasitism occurred in the laboratory or in the field. The chamber of the affected pupa seemed crudely done, mainly built with plant fiber and missing a more consistent humus wall.

Pupal morphology
Generally, pupae of Scarabaeidae are of uniform appearance, mainly differing by adult characters such as horns. Usually they do not have pupa-specific ornamentations or setae for chaetotaxy analysis unlike pupae of other beetle families which might be the reason for the poor attention that scarabaeoid pupae have gained in the past (see Costa et al. 1988;Costa and Ide 2008).
Several descriptions of pupae from Dynastinae tribes, such as Cyclocephalini, Phileurini, Oryctini point out the thin golden setae present at the apex of 9 th abdominal segment but rarely mention the shape of the 9 th tergite and the genital ampulla (Phileurini: Vanin et al. 1983;Morelli [1991] Ratcliffe and Chalumeau 1980;Costa et al. 1988;Morelli 1997;Alvarez Castillo et al. 1998;Neita-Moreno and Orozco 2009;Pardo-Locarno et al. 2009), characters which seem to vary within Scarabaeidae. Our examination of pupae from Phileurini and Oryctini revealed that they can be distinguished from that of P. felscheanus based on the form of the 9th tergite and genital ampulla and the distribution and size composition of those apical setae.

Comparison with other scarabaeid pupae
In a pupa of Homophileurus luederwaldti examined (length: 25.3 mm, largest width: 13.0 mm; Phileurini) the setation on the apex differs from that of P. felscheanus by covering a much wider area, until the middle of the ventral part of t9, and the setae also extend slightly towards the dorsal side. The apical notch is more open than it is in P. felscheanus. The genital ampulla region was damaged, and could not be analysed. In a male pupa of Trioplus cylindricus (length: 22.0 mm, largest width: 9.3 mm; Phileurini), the setae on the apical tergite are similar to the pupa of H. luederwaldti. They cover a larger area, spreading laterally towards the dorsal region, but almost do not change in density and sizes as in H. luederwaldti. The rounded structure of the genital ampulla is wider posteriorly and narrowed from middle to base, thus not rounded or elliptical. There are two longitudinal marks, one smaller anteriorly and the longer posteriorly. In a male pupa of Strategus validus (length: 60 mm, largest width: 27 mm; Oryctini), the rounded structure of the genital ampulla is elliptical with one longitudinal mark at the apex and thus, similar to that of P. felscheanus. Tergite 9 has a transversal depression laterally (not present in P. felscheanus); pubescence of uniform, small setae covering a wide area throughout the segment, ventrally covering almost the whole area extending laterally unto the edge of the lateral depression; and an apical notch connected mesally by V-shaped fold (absent in P. felscheanus).
Furthermore, in Hemiphileurus elbitae Neita-Moreno and Ratcliffe 2010 (Phileurini), the apex of t9 appears to be more densely setose over a larger area, and with longer setae (Neita-Moreno and Ratcliffe 2010) than in P. felscheanus. The apices of male and female pupa of Homophileurus tricuspis Prell, 1914 (Phileurini) have setae that extend significantly more sidewards (Neita-Moreno and Ratcliffe 2011) than in P. felscheanus. The pupa of Phileurus affinis Burmeister, 1847 (Phileurini) has t9 covered with small and abundant setae on the apex, but there is no detailed illustration to compare (Morelli [1991]). The female pupa of Aspidolea singularis Bates, 1888 (Cyclocephalini) has the setae on t9 similarly distributed as those of P. felscheanus, but dorsally and laterally the shape of t9 is different. In the former, it is almost as long as wide, and in the latter it is two times wider than long. In addition t8 is more sharpened medially and projected towards the apex, whereas it is rounded in P. felscheanus. The pupa of Ancognatha ustulata Burmeister, 1847 (Cyclocephalini) has setae widely distributed ventrally on t9 (Neita-Moreno and Morón 2008).
Even though the knowledge of pupae is much more limited than that of larvae, pupae do have some exclusive characters such as setae, spurs, tubercles and modified spiracles (Costa and Ide 2008). As demonstrated here, the apex of the abdomen, namely the shape of t9, genital ampulla, and setal pattern could be a source of important characters for distinguishing between subfamilies, tribes, and genera of Scarabaeidae and perhaps could even be used in systematic and phylogenetic analyses, but need to be much more comprehensively studied. Altogether, we consider the data from pupal morphology presented here to be too limited to contribute to the discussion of the systematic placement of P. felscheanus.

Life history
Dynastinae larvae frequently feed on decaying plant matter, especially wood (Ritcher 1958;Carlson 1991;Ratcliffe et al. 2002), although food sources vary between species and include roots of living plants, especially monocots, and general organic matter (Ritcher 1958;Gassen 1989). Lourenção et al. (1999) reported larvae of Strategus Hope, 1837 (Oryctini) as a pest of Arecaceae feeding on live plant tissue. Larvae of Oryctini and Cyclocephalini are often associated with feeding on live plant tissue, unlike those in the tribes Phileurini and Dynastini that are known to feed on decaying wood or humus associated with decaying vegetable matter. Thus P. felscheanus would be the first known exception in the Phileurini tribe. Krell et al. (2002) reported larvae of a species of Cetoniinae feeding on living bromeliad tissue and already referred to an unidentified large dynastine larva collected by Lüderwaldt (1915) in bromeliads at Colonia Hansa, Blumenau, in Santa Catarina state, Brazil. Nevertheless, this is the first report of identified larvae of Dynastinae feeding on living leaf tissue (Fig. 32) and the only species of Dynastinae known so far whose larvae develop in bromeliad rosettes.
Considering the place where we found the younger larvae, females most likely lay eggs into the external leaf axils that are rich in decaying plant matter and humus, and usually dryer than the water-filled central rosette. Our observations suggest that the larvae initially feed on dead decaying vegetal matter and then migrate into the center of the rosette to feed on the white basal part of living leaves. It is during this second phase that the larvae gain most weight.
The construction of pupal chambers in Dynastinae species is poorly documented. Larvae of Phileurus hospes Burmeister, 1847 built their pupal chamber only with humus from the decaying tree trunk in which the larvae were found (F. F. Albertoni pers. obs.). The termitophilous species Homophileurus luederwaldti, Actinobolus tribolus Luederwaldt, 1910 and Actinobolus radians Westwood, 1841 used their own faeces and soil from the termite nest to build their pupal chambers (Luederwaldt 1911). Homophileurus tricuspis Prell, 1914 seemed to make a hole in the termite nest wall where the larvae delimited their pupal chamber (Neita-Moreno and Ratcliffe 2011). Trioplus cylindricus larvae built their pupal chamber excavating the wood or using fragments of wood (Vanin et al. 1983). As is the case with P. felscheanus, these larvae seem to have opportunist behaviour by using the substrate next to them. The location, size, and the material used by P. felscheanus for building its pupal chamber (Fig. 34) resembles those of another bromeliad associated insect, the lepidopteran Geyeria decussata (Castniidae) (Albertoni et al. 2012) except for the fact that the beetle larva does not use silk.
According to Ritcher (1958), most Scarabaeidae push the last larval exuvia to the rear of pupal chamber during pupation, but in Rutelinae and most Dynastinae the last larval exuvia splits longitudinally along the middorsal line and the pupa stays inside the larval exuvia. However, in a few genera such as Oryctes Hellwig, 1798 and Strategus Hope, 1837 (both Oryctini) the larval exuvia is pushed to the rear of the pupal chamber as was observed with Platyphileurus (Fig. 34).
Among the South American Coleoptera that are hosts of Tachinidae, Chrysomelidae and Scarabaeidae are the families with the highest numbers of parasitized species (Guimarães 1977), but none of the 7 genera of Goniini listed in Guimarães's catalogue parasitized Coleoptera. Neither was the monotypic genus Platyphileurus Ohaus, 1910 registered as a host, nor was Mystacella van der Wulp, 1890 (Figs 36-37) registered as parasite for any beetle species. In North America, Mystacella spp. was recorded as parasitizing Lepidoptera of the families Arctiidae and Noctuidae and M. chrysoprocta (Weidemann, 1830) parasitizing Scarabaeidae of the genus Xyloryctes Hope, 1837 (Arnaud 1978). Thus, our observations constitute a new host-parasitoid association.

Transfer of Platyphileurus from Phileurini to Oryctini on the basis of larval and adult characters
Since the species studied here was not only described as two species, P. felscheanus and S. jelineki, but also as belonging to different tribes, namely Phileurini and Cyclocephalini, it is appropriate to explore whether larval characters can contribute to resolving its tribal classification. Antennal sensory spots: Larvae of Cyclocephalini are characterized by 2, 3, or 4 sensory spots on the last antennomere plus an apical one (Ritcher 1944;Remedi de Gavotto 1964;Morelli and Alzugaray 1994;Ramírez-Salinas et al. 2004;Neita-Moreno et al. 2007;Bezerra de Souza et al. 2014). With 12 to 13 dorsal and 13 to 15 ventral sensorial spots on the apical antennomere, P. felscheanus would be the only known exception in this tribe.
Other larval characters: According to Neita-Moreno and Orozco (2009), Oryctini larvae can be diagnosed by the combination of the following characters: cranium densely punctate, dark reddish brown; maxillary stridulatory teeth truncate; antennomere 4 with 2-15 dorsal sensory spots; tarsal claws with 2-4 long, stout setae; raster without palidia or septula. Platyphileurus larvae show all those characters. According to the larval characters this genus could belong to Oryctini, or possibly Dynastini (but these having rather rounded stridulatory teeth), but it is unlikely to belong to Cyclocephalini, Phileurini, or Pentodontini.
Adult characters: To exclude the possibility of Platyphileurus belonging to Oryctini, the adult mouthparts were examined. The mentum (labium) of Platyphileurus is triangular, basally broad, tightly tapered to a blunt, rounded, thin and slightly protruding tip (Fig. 29). The basis of the labial palps is almost visible, being slightly covered by the margin of the mentum only. This is different from the form of the mentum diagnostic for Phileurini, being broad and covering the basis of the labial palps completely. In fact, the mentum of Platyphileurus resembles broad menta of Oryctini or Pentodontini.
Problems with the current tribal classification: In the current, typological classification, Oryctini are separated from Pentodontini by one variable adult character, the apex of the hind tibia, of which Ratcliffe and Cave (2006: 189) noted: "We remain concerned that this single, sometimes variable (or transitional) character used to separate taxa at the tribal level is not reliable." Likewise, Dynastini are separated from Oryctini by only two adult characters: a cylindrical first tarsomere of the hind legs (triangularly dilated in Oryctini and other tribes) and anterior legs in males more or less prolonged (no such dimorphism in Oryctini and other tribes) (Endrődi 1985;Ratcliffe and Cave 2006). Although Dynastini could well be monophyletic characterized by those two potential autapomorphies, it would leave the Oryctini without an autapomorphy in the current classification since the character separating it from Pentodontini, the crenulated or denticulate metatibial apex, is also present in Dynastini and possibly a synapomorphy of Oryctini and Dynastini. Without an autapomorphy, Oryctini cannot be diagnosed as a monophylum (cf. Hennig 1982: 93;Wheeler 2012: 41). Being entirely typological, the current tribal classification is likely to contain paraphyletic or even polyphyletic tribes.
In a caryological study Dutrillaux et al. (2013) found that Augosoma Burmeister, 1847, currently in Dynastini, might be closer to Oryctes Hellwig, 1798 (Oryctini) than to other Dynastini. The recent cladistic analysis of Dynastini by Rowland and Miller (2012) proposes Augosoma to be sister to the remaining analysed genera of the subtribe Dynastina, indicating its early branching within the Dynastini. This analysis cannot help clarifying the relationship between Dynastini and Oryctini (or Oryctini+Pentodontini) since representatives of the latter were not included. The relationship between Oryctini and Dynastini remains unresolved, but the recent studies indicate that Oryctini could be paraphyletic in respect to a -possibly polyphyletic -Dynastini.
Tribal placement of Platyphileurus: The apex of the hind tibia in Platyphileurus (Fig. 30), being (weakly) dentate and not truncate, would place this genus in Oryctini or Dynastini, not in Pentodontini in the current sense. This is supported by the number of sensilla on the larval antenna and the shape of the left larval mandible. The slightly broader first tarsomere of the hind legs, together with the small body size being unusual for Dynastini, indicates that it rather belongs to Oryctini than to Dynastini in the current sense. The shape of the adult mentum, the large number of sensilla on the larval antenna, and the strong S4 of the left larval mandible are not found in Phi-leurini. We accordingly propose the transfer of Platyphileurus from Phileurini to Oryctini. We consider the flat body of Platyphileurus a convergence with the body shape of Phileurini, likely related to the unique habitat among the tight bromeliad leaves, where the larva pupates and that enable adults hiding between their leaves.