Comparative morphology of the mouthparts of the megadiverse South African monkey beetles (Scarabaeidae: Hopliini): feeding adaptations and guild structure

Light microscope

Heads were removed from the body and imbedded in a melted wax-rosin mixture with a soldering iron and insect pins. Mouthparts were dissected using a stereo microscope and embedded in Polyvinyllactophenol on a microscopic slide. After drying for 48 hours specimen were sealed with nail polish to prevent them from draining and to avoid infiltration of air. Images of different focal planes were taken with an Olympus CX41 light microscope equipped with a digital Olympus E330 camera. Micrographs were stitched using Helicon Focus software (Version 3.10) and processed with Adobe Photoshop CS5 (Version 12.0). For each of the 18 collected species, between 3 to 11 specimens were investigated in detail.

The contents of the anterior gut were extracted using a stereo microscope and microscope slides were prepared in Polyvinyllactophenol. After air drying, the gut contents were examined using a light microscope (Olympus CX41).

Scanning Electron Microscope (SEM)

Heads and dissected mouthparts were dehydrated in absolute ethanol and submerged in Hexamethyldisilazan. After air drying overnight, specimens were attached to an aluminium stub with graphite adhesive tape and sputter-coated with gold (240 s, Agar sputtercoater B3740). Micrographs were taken from 1 to 3 specimens of selected representatives with a Philips XL 20 SEM and a Philips XL 30 ESEM (Philips, Amsterdam, Netherlands).

Statistical analysis

Data analyses were conducted using Correspondence Analysis (CA) with the statistical computing software R 3.2.2 (R Core Team, 2015). This analysis allows trait and feeding guild to be statistically related. Results were visualized with a balloon plot (function balloonplot) in the “gplots” package (Warnes et al., 2015), which uses a contingency table based on a character matrix of mouthpart traits and graphically shows the relative magnitude of these traits to different feeding guilds. The contingency table was built using the package “factoMineR” (Husson et al., 2015) and the function “devtools” within the “factoextra” package (Kassambara, 2015).

General mouthpart morphology

All investigated species displayed prognathous mouthparts that consisted of the labrum with the epipharynx, paired mandibles and maxillae, as well as the labium with the hypopharynx. The maxillary and labial palps were found to be similar in size and shape across all studied species. Mouthparts were found to differ only in size, but not in form between sexes. Consequently, no conclusions about differences in the feeding preferences between males and females were drawn.

Labrum and epipharynx

The labrum, together with the clypeus, formed the cibarial roof. Overall, the labrum of all investigated species was a sclerotized, more or less heart-shaped structure, equipped with long hairs, particularly on the concave distal edge (Figs. 1A, 1E, 1I, 2A and 2E). The latter also showed strong sclerotization with lateral bulges, and ventrally a v-shaped hair crest composed of different numbers of setae was seen connected to the membranous, haired epipharynx. In contrast to the other investigated species, the labrum of Anisonyx ursus, Pachycnema grassipens and P. calcarata was weakly sclerotized. Further, in the case of A. ursus, it was elongated and slender in appearance with a u-shaped indentation on the distal edge (Fig. 1I).

Mandibles

The paired, stout mandibles across all the hopliine species examined were mostly rectangular in shape. Each mandible consisted of a proximal, heavily sclerotized mola with a membranous pilose postmola, a membranous lacinia mobilis and a distal blade-like incisivus (Figs. 1B, 1F, 1J, 2B, 2F and 3A–3H). The strongly sclerotized mola was a socket-shaped structure with a grinding surface equipped with several rows of small pointed teeth. In contrast to all other investigated species, the left and right mola of Pachycnema flavolineata differed concerning the surface structure. On the surface of the right mola, a sclerotized, curved ledge could be found; the left molar displayed a corresponding ridge (Figs. 3D and 3E). Heterochelus pickeri, contrasted strongly with all other species in that it showed a significantly reduced molar (Fig. 3F), resulting in a slim, triangular shaped mandible (Fig. 2B).

The main part of the mandible for all hopliine species was formed by the convex lacinia mobilis and this more or less pilous lobe reached from the mola to the incisivus. In some species (e.g. Mauromecistoplia nieuwoudtvillensis) the lacinia mobilis was covered with rows of short, tooth-shaped structures (Fig. 3G). Clania species, Congella species, Lepisia species and Scelophysa scheffoldi possessed a single, lateral tooth on the lacinia mobilis. In Chasme decora and Lepithrix sp. only a few hairs were found on the lacinia mobilis, but a conspicuously hooked tooth protruded at the base (Figs. 1F and 3B). In A. ursus, and in all three Pachycnema species, the mandible was equipped with a conspicuously large, square-shaped and soft lacinia mobilis (Figs. 1J and 2F).

Across most species the smooth, rounded incisivus was sclerotized, usually lacked a cutting edge and protruded distally from the heavily sclerotized basal joint region. In Congella sp., Kobousa gentilis and M. nieuwoudtvillensis a cutting edge was, however, evident, formed by the bent apex of the incisivus (Figs. 1B and 3H).

Content of the alimentary tract

Flower tissue was found in the guts of both Clania species, Chasme sp., Congella sp., Dolichiomicroscelis gracilis, Kubousa gentilis, both Lepisia species, Scelophysa scheffoldi and M. nieuwoudtvillensis. In addition, various amounts of intact pollen grains could be found on the mouthparts and in the guts of each of these species. All dissected specimens of C. decora, P. crassipes and P. flavolineata had large amounts of intact pollen grains in their guts, with some grains found on the mouthparts as well. No pollen grains could be found in the gut of A. ursus and P. calcarata; although a colourless or orange substance was found in the foreguts of both species. While no pollen grains were found in the dissected gut of H. pickeri, small amounts were found on the mouthparts.

Correspondence between characters and feeding guilds

The Correspondence Analyses (CA) indicated a significant association between mouthpart characters and feeding guilds (r = 0.63; Chi-square of independence, χ²(22) = 43.48, p = 0.004). The balloon plot (Fig. 4) displays the relative magnitude of characters to different feeding guilds of the investigated species, and the CA factor map (Fig. 5) gives a measure of similarity between mouthpart characters (blue points) and feeding guilds (red triangles). Elongated mouthpart structures (mel, cse and gae) and ligulae of the labium (lal) corresponded with nectar feeders. In addition, pollen and floral tissue feeders corresponded closely with a prominent, toothed mola (pmo, mot) and a large, densely bristled (lml, lmb) and sometimes toothed lacinia mobilis (lmt) as well as a bristled galea (gab). Incisivi with cutting edges (ice) and toothed galeae (gat) were only found in the floral tissue feeding guild. This indicated that, although pollen- and floral-tissue feeders shared some characters, the three feeding guilds could be distinguished from each other by reference to their mouthpart morphology.

General hopliine mouthpart structure

Both the labrum and labium, with the haired epipharynx and hypopharynx, respectively, serve as the dorsal and ventral cover of the cibarium and seal the oral cavity. While the maxillae represent the primary nutritional uptake organs in scarab beetles, the mandibles are responsible for the processing of food within the cibarium (Karolyi, Gorb & Krenn, 2009). In hopliines, the labrum and labium showed fewer modifications among the feeding groups than the maxilla and the mandible.

Pèringuey (1902) characterized the maxilla of hopliines as bilobated (inner lobe, the lacinia; and outer lobe, the galea), whereby the lacinia is often indistinct or absent and is always immovable (see also Williams, 1938). These first descriptions of the maxilla coincide broadly with those found in the present study. The inner lobe, the lacinia, which usually represents the sclerotized and robust part of the maxilla in scarab beetles, was reduced in hopliines. In the hopliine feeding groups that target nectar and pollen, the lacinia, although reduced, was found to be covered with pads of hairs. These are most likely adaptations towards mopping-up nectar and sweeping-up pollen. The elongated, basal cardo possibly allows an increased movability in protruding the maxilla to gather pollen and transport it to the mouth (Schremmer, 1961; Fuchs, 1974; Johnson & Nicolson, 2001; Krenn, Plant & Szucsich, 2005). As such, we here show that hopliines show several modifications of the maxilla towards specialised and different feeding strategies.

Classification of hopliine feeding groups

Based on their mouthpart morphology and gut content, the investigated species were classified into three different feeding groups: floral/plant tissue feeders; pollen and nectar feeders (Table 2; Figs. 1 and 2) and one species that we could not place accurately, but which may have an intermediate feeding strategy utilizing floral tissue, pollen and perhaps nectar (Fig. 2). Although the gut content analyses helped guide the understanding of mouthpart specialisations towards feeding on a particular floral resource, greater sampling of individuals would be required to confirm this. Our results were mostly based on a few random specimens per species. Future studies should more systematically sample individuals from all host plant species being used. Additionally, sampling of individuals of a species from different plant communities and habitats would confirm if a particular floral resource is targeted irrespective of the changing plant species in a community. As such, our gut content analyses should be treated with some caution and considered as preliminary at this stage.

Feeding on floral and foliage tissues

Ten out of the 18 investigated hopliine species were classified as feeding on flower tissue, and one species on foliage tissue (Table 2). Hopliines characterised as floral or foliage tissue feeders displayed a similar mouthpart morphology characterized by teeth and cutting edges on the maxillae and mandibles. Picker & Midgley (1996; see also Midgley, 1993) regarded such species as being destructive to their host flowers. Several differences in the morphology of the mandibles, especially in the shape of the incisivus, were found in this group and not in others. The sclerotized cutting edges on the mandibles, together with the toothed galea, are considered as adaptations towards floral tissue feeding. The rows of short, tooth-shaped hairs as well as the single, lateral tooth found in some species (e.g. Clania) on the lacinia mobilis of the mandible are possibly used to grate floral or foliage tissue, whereas the sclerotized molar is most likely used to grind floral particles. Hopliines characterized as flower and foliage feeders were found to be the only ones with cellulose present in their gut; although not all specimens in this group contained cellulose in their guts.

Pollen-feeding

The challenges confronting insect pollen feeders are the handling and removal of pollen grains and gaining access to the nutrients encapsulated in the hard and highly resistant exine. Pollen is a valuable food source for flower visiting insects since it contains various lipids, carbohydrates and proteins (Roulston & Cane, 2000; Johnson & Nicolson, 2001; Cook, Murray & Williams, 2004). Characteristic adaptations of pollen feeding in Coleoptera include elongated maxillae with variously shaped combs and hairs that function as pollen harvesting devices (Fuchs, 1974). In Coleoptera, the mandibles are particularly adapted for pollen manipulation and consist largely of the lacinia mobilis (or prostheca), a soft and usually bristly lobe, and the postmola, which together knead the pollen (Schremmer, 1961; Fuchs, 1974; Nel & Scholtz, 1990). The pollen grains are conveyed into the actual mouth opening by coordinated movements of the mouthparts (Fuchs, 1974; Karolyi, Gorb & Krenn, 2009).

Only three species of hopliine (Table 2) based on the general mouthpart characteristics of pollen feeders as described above, could be classified as specialised pollen-feeders. Adaptations for pollen-feeding in these species can be seen in several modifications of the mandible and the maxilla. The lobe-like lacinia mobilis, with its hooked tooth, most likely functions as a pollen collecting device inside the cibarium, as seen in other scarab beetles, where during movement of the mandibles, the lacinia mobilis acts as a scraper, wiping pollen grains to the molar part. The two hopliine species in this specialised pollen-feeding group possessed a prominent mola with numerous rows of teeth that may be used for perforating pollen grains.

In pollen-feeding Scarabaeidae, the distal part of the maxilla, i.e., the haired galea, is considered as the primary organ for pollen uptake, and the sclerotized hairs are expected to function as an ideal pollen brush (Karolyi, Gorb & Krenn, 2009). Although the lacinia was found to be reduced in these specialised hopliine pollen-feeders, the hair comb of the lacinia most likely is used to convey pollen grains to the mandibles. In nectar- and pollen-feeding Meloidae, the haired ligulae are considered specialised organs for harvesting pollen and nectar (Wilhelmi & Krenn, 2012). Therefore, the forward pointing ligulae on the labium of pollen-feeding hopliines are hypothesised to have the same function.

Since most of the pollen grains found in the gut of pollen-feeding species were intact, it can be suggested that the broad molae of the mandibles are most likely not used to crack open pollen. South African protea beetles (Cetoniinae) and some monkey beetles are known to use osmotic damaging of the pollen wall inside the gut, enabling digestive enzymes to enter via the pores (Johnson & Nicolson, 2001; see Karolyi, Gorb & Krenn, 2009 for summary of various mechanisms of pollen digestion in Coleoptera). The teeth found on the molar of investigated species might therefore be used primarily to pierce pollen grains, enabling digestive enzymes to enter within the gut.

Nectar-feeding

Sugar concentrations in nectar mostly range between 15–60%, depending on the plant species (Daniel, Kingsolver & Meyhöfer, 1989). In addition, nectar contains small amounts of amino acids, proteins, organic acids, phosphates, vitamins and enzymes (Baker & Baker, 1973; Kevan & Baker, 1983; Nicolson, 2007). Sugars in nectar provide the primary energy source for fast flying flower visitors that have high energetic requirements, such as hummingbirds and various insects. Although, several south African Iridaceae offer nectar to insects with elongated mouthparts, like long-proboscid Diptera (Goldblatt & Manning, 2000; Johnson, 2004; Karolyi et al., 2012; Karolyi et al., 2013; Karolyi et al., 2014), some plant species have adapted their flower morphology towards attracting hopliines with secreted nectar (Goldblatt & Manning, 2011). Picker & Midgley (1996) classed species of Peritrichia and Anisonyx as nectar feeders. However, no species of Peritrichia examined in this study were determined to be nectar feeders; only A. ursus and Pachycnema calcarata and P. crassipes were identified as specialised nectar feeders (Table 2). Picker & Midgley (1996) classed species of Pachycnema as pollen and floral tissue feeders based on their general behaviour of embedding deeply within the capitulum of flowers.

The three investigated species characterised as nectar feeders possessed long and slender mouthparts that lacked any cutting edges that could be used for damaging their host plants (Figs. 1I–1L). The labrum differed in shape and was rather weakly sclerotized compared to the flower and pollen-feeding species. Furthermore, the broad shape and soft texture of the lacinia mobilis suggests the ability to sweep nectar into the mouth within the cibarium. The long and slender maxillae, with the hirsute and heavy sclerotized galea, are most likely adapted for probing into flowers and for mopping-up nectar. The long and slender cardo found in nectar feeding hopliine species presumably allows protrusion and a higher movability of the distal maxilla parts during food uptake. In addition, the elongated labium, together with the elongated, hirsute ligulae, was also considered to be an adaptation towards nectar feeding in these three hopliine species.

Florivory, pollen or nectar feeding, or a mixed cocktail?

Some of the investigated species showed mouthpart adaptations for a dual diet, suggesting that these species may feed on more than one type of floral food source. In particular, Heterochelus pickeri and Pachycnema flavolineata could not clearly be categorized into a specific feeding group because their mouthparts displayed adaptations to several different food sources. Although the maxillae appeared quite similar between the two species, the mandibles differed significantly.

Picker & Midgley (1996) described Heterochelus species as being destructive to their host plant´s flowers through their feeding behaviour. Concerning the mandibles of H. pickeri, this seems contradictory since they possessed no cutting edges and the reduced mola was unsuitable to grind or crush any food particles. Further, the lacinia mobilis, together with the postmola, appear to function rather as a conveying structure for pollen and/or nectar, like in European Cetonia beetles (Karolyi, Gorb & Krenn, 2009). In addition, the galea brushes most likely serve as effective pollen up-taking organs. From its position in the head capsule, the toothed galea seemed better suited towards combing pollen grains, rather than for feeding on floral tissue. Compared to the floral tissue feeding species, the galea teeth of both H. pickeri and P. flavolineata were rather short, possibly functioning more as a pollen rasp, grating pollen from the anthers. Although the mandible of H. pickeri appears not to be adapted to grind pollen, beetles may ingest pollen intact and digest these via an osmotic gradient in the gut, as described for other hopliines (Johnson & Nicolson, 2001).

In P. flavolineata, the shape of the mandible suggests that this species feeds on pollen. Modifications of the mandible, such as a large, hairy lacinia mobilis and postmola, as well as a strongly sclerotized mola, have been reported for beetles feeding on pollen and nectar (Krenn, Plant & Szucsich, 2005). In addition, the surface of the mola might function like a tongue and groove joint, interlocking the mandibles during feeding. This might enhance the force that can be applied to crush pollen grains. The combination of the toothed, but also hairy galea of the maxillae of P. flavolineata, together with the well-developed ligulae of the labium represents a combination of traits for pollen uptake as well as floral tissue feeding. This suggests a twofold diet for this species, although, based on the overall shape of the mouthparts, it can be concluded that pollen possibly represents the primary food source. This was confirmed by gut content analysis, where several specimens were found to have numerous Asteraceae pollen grains in the gut.

Relating mouthpart adaptations to flower visiting behaviour

Although some matches between monkey beetle pollinator guild placements by Picker & Midgley (1996) could be confirmed with the results of this study, several of the former classifications do not conform to current findings. According to the mouthpart structure of the species examined here, species that have previously been classified as pollen and/or nectar feeders should now be considered as primarily floral tissue feeders (e.g. Clania species), whereas floral tissue and/or pollen feeders should now be regarded more as nectar and pollen feeders (e.g. Pachycnema species). Most of the feeding preferences have been derived from their flower visiting behaviour. However, the results presented here suggests that not all flower visiting behaviours are linked with feeding and selection on the primary food source.

Beetles belonging to the embedding group mostly visit Asteraceae and Aizoaceae and are typically found buried deeply into flowers (Picker & Midgley, 1996). Because of this behaviour, species of this group have been considered destructive to their host flowers and their role as pollinators have been questioned. However, according to their mouthpart morphology, only Scelophysa scheffoldi can be regarded as a strictly floral tissue feeder, while the other species within this group are nectar and/or pollen feeders. The embedding behaviour may therefore be due to reasons other than feeding on floral tissue, e.g. possibly related to predator avoidance (Ryan, Colville & Picker, 2013; see also Colville, 2009 for additional hypotheses associated with the embedding behaviour in hopliines). Consequently, some hopliine species can be destructive to their host flowers because of their burrowing and not necessarily their feeding behaviour.

Because of the mismatches found between pollinator guilds established by Picker & Midgley (1996) and the feeding groups described in this study, current guild definitions of hopliines need refining so that mouthpart morphology is also considered.