Cassida sphaerula Boheman, 1854

Natural history of Cassida sphaerula Boheman, 1854 View in CoL

Field observations over almost one year revealed the beetle’s cycle of activity. Our observations began at the end of one breeding season. From early summer (27.XII.2021) till early autumn (mid-May 2022) no adults or larvae were seen. Then the new breeding season began in late autumn. The first sign of beetle activity is the small ‘windows’ chewed on the plant leaves in autumn (late May); then the larvae can be seen on the underside. Egg-laying begins in May, with much larval activity by mid-June, when minimum temperatures are around 35.6°F (2°C). Up to 8 egg cases have been found on a single leaf. Thus, we believe this species is an autumn/winter breeder.

Egg cases (n =4) ( Fig. 16–18 View Figures 16–18 ). Oothecae are deposited on the venter of the leaf ( Fig. 16 View Figures 16–18 ) in apparently random areas between veins. We observed a maximum of eight oothecae per leaf. The ootheca lies flattened on the long axis, shallowly tucked into the leaf surface as there is a slight depression under each one; it is not stalked, suspended or protuberant from the leaf surface, as in some other Cassidinae . Oviposition was not observed so it is unclear how the female may prepare a site before depositing her eggs (see Müller and Rosenberger (2006) for possible oviposition sequences in Chrysomelidae ). The ootheca, secreted by colleterial glands ( Hinton 1981; Gillot 2002), comprises a thin opaque outer laminate membrane that appears shiny and dark brown ( Fig. 17–18 View Figures 16–18 ). The few enclosed eggs (less than 5) are cream-colored ( Fig. 17–18 View Figures 16–18 ), dorsoventrally compressed (lying flattened on leaf), and elongate-oval shaped. The ootheca lacks any additional coverings, no fecal or chewed plant material. Egg hatch. One ootheca was collected on 11.X.2021 and three larvae hatched on 18.X.2021, confirming that more than one egg is oviposited at a time. We did not observe how the larvae exited the egg case, but we found the ootheca roughly torn at one end and left behind, therefore not eaten by the neonate.

Larva (n =20; Fig. 13–14 View Figures 9–15 , 16 View Figures 16–18 , 19–25). The larvae are solitary, not apparently gregarious, but may be found mixed with others of different stages in a dense situation, even feeding side by side with scoli (lateral projections) in contact. They do not respond to disturbance by moving into groups or with coordinated cycloalexic (ring) defense where larvae move into a tight, somewhat circular, group and all flex the shield in unison (see Jolivet et al. 1990).

Larvae are found mostly on the venter of the leaves. Instar I (n =3; Fig. 19–20) are tear-dropped shaped, about 2 mm long X 1 mm at maximum width (across pronotum). The body is tan-colored. The paired caudal processes ( Fig. 21, 24–25 View Figures 21–25 ; = supra-anal processes, urogomphi) are almost half as long as the body. Older larvae ( Fig. 23 View Figures 21–25 ) are creamy yellow and with a dark brown central area; their cuticle is almost transparent, and the internal organs are somewhat visible (internal movements are easily seen). The scoli pattern ( Fig. 21–22 View Figures 21–25 ) and caudal processes processes are similar between instars and fit with Świętojańska’s (2009: 74) generalized Cassida larvae having an ovoid dorso-ventrally flattened body with 16 pairs of lateral scoli.

Exuvio-fecal shield. The larval shield is initiated in Instar 1 (Fig. 19) shortly after it initiates feeding. This shield is comprised only of larval feces that is applied to the caudal processes by the muscular telescoped anus. The shield grows into an elongate mass on the larva’s paired caudal processes (Fig. 20). The shield can appear dry ( Fig. 22 View Figures 21–25 ) or wet ( Fig. 23 View Figures 21–25 ) and the telescoped anus periodically applies a dark wet droplet (see Fig. 13 View Figures 9–15 ) to the shield. Dissections of shields reveal a fan ( Fig. 22 View Figures 21–25 ) or pyramidal shape ( Fig. 23 View Figures 21–25 ) with a central scaffold of stacked, nested exuviae and all entirely covered in dry or moist feces. The exuviae are not easily discerned in intact shields (e.g., Fig. 23 View Figures 21–25 ). In dissected shields ( Fig. 24–25 View Figures 21–25 ), the feces are abraded to reveal the stack of exuviae; each exuviae can be individually teased off to show the caudal processes of older instars. The larvae continue to build, applying fresh feces and wet droplets (note wet appearance in Fig. 24 View Figures 21–25 ).

Pupa (n =10; Fig. 26–27 View Figures 26–31 ). The pre-pupal stage is typically when the mature larvae ceases feeding, become sedentary and fixes its abdomen to the substrate. Five young larvae were followed (three from egg hatch) to adulthood; three pupated; pupation lasted nine days, 15 days, and 20 days. Six mixed-age pupae/pre-pupae were placed in a container on 28.VIII.2021 and the first adult appeared on 7.X.2021 (9 days); two of these pupae failed and four adults were reared. Thus, pupation (n=7) ranges from 9–20 days. No parasitoids emerged from these laboratory pupae.

Pupae ( Fig. 26–27 View Figures 26–31 ) are ~ 9 mm long, solitary, affixed by their abdomen to the leaf venter, never on the upper part. There is seldom more than one pupa per leaf. The pupa is tan-colored, and the body is ovoid and dorsoventrally flattened. Only the abdominal segments have lateral scoli.

The pupa of C. sphaerula shows two types of shields. It may retain the final exuviae ( Fig. 26 View Figures 26–31 ) and the former larval shield may be found discarded nearby or the pupa may retain the entire shield structure of the 5 th instar larva (exuviae I–IV and their fecal matter) ( Fig. 27 View Figures 26–31 ). As far as we know, this is the first observation of such flexibility in shield retention in cassidine pupal shields. After the adults have emerged, the pupal exuviae remains attached to the plant for a long time, with or without the fecal shield. A few adults seem to have some difficulty eclosing, taking longer and struggling to exit the exuviae, but these adults eventually became hardened and moved away.

Adult (n =30; Fig. 28–30 View Figures 26–31 ). These are ~ 9 mm long (along midline, head to posterior margin) by 4–5mm at their maximum width (across pronotum). The dorsum is generally pale green in color but can vary from translucent straw to a deep green. They were observed as early as 30 August (reared) and 23 September (wild) and are generally solitary. During the observation period, the habitat experienced a frost (late August) and the beetles remained sluggish but resumed activity as temperatures rose.

Color/pattern variations. Newly eclosed or teneral adults are straw (pale-brown) colored and the mature hardened adults are green. We only observed mating pairs of green individuals. As adults age, some acquire permanent circular blackish marks in different locations of the elytra ( Fig. 30 View Figures 26–31 ), but we did not detect marked color polymorphism as in some Cassidini ( Simon Thomas 1964; Verma and Kalaichelvan 2004).

Courtship and mating ( Fig. 13 View Figures 9–15 , 29 View Figures 26–31 ). Mating pairs were first observed on 24.VIII.2021, as the frost season ended, and the region transitioned to spring. Courtship was not observed by the many mating pairs found, but pairs in copula were noted. Mature green adults exhibited no rapid (a few seconds) color changes (with temporary black spots or to golden or straw colors) as documented for some Cassidinae during mating or when disturbed ( Barrows 1979).

Dormancy. Beetle activity ceased as the summer peaked and it is unclear where they hide. The host plants do not lose leaves in winter, suggesting that the beetles can have a steady food supply, further supporting them as a good biocontrol agent. We continue observations in 2022 but have not determined if beetles pass the winter hidden under stones or in dead vegetable matter, as they tend to do in the Natal area (Heron, pers. obs.).

Feeding patterns ( Fig. 11–12, 14–16 View Figures 9–15 View Figures 16–18 , 28 View Figures 26–31 ) of C. sphaerula . Larvae and adults feed in similar ways, which creates a distinct pattern of craters on the venter of the leaf, each crater with the cuticle rolled to one side ( Fig. 15 View Figures 9–15 ). The craters of instar 1 are small (Fig. 19–20); older larvae and adults make craters up to 4 mm long. The craters are hollowed out by feeding and are irregularly shaped (hemispherical, ovoid, rounded). They have a deep basin, with the rolled ventral cuticle forming a thickened margin on one side. The dorsal cuticle of the leaf remains intact ( Fig. 11 View Figures 9–15 ), with a window-pane pattern. The mid-rib and secondary veins are not eaten but the leaves are intact dorsally and do not have a skeletonized appearance.

Larvae start feeding shortly after hatching. Their feeding exhibits a stereotyped repertoire. The site is prepared by eating most of the trichomes (Video 1). The first cut of the ventral cuticle (including the leaf ’s epicuticular wax layer) is made by a series of bites that create an arc-shaped cut, about the same size of the pronotum. The larva starts feeding on spongy mesophyll, and its head action pushes the cuticle layer, rolling it over and ventrad. As the larvae feed on the exposed mesophyll, a crater forms, deepens, and enlarges ventrad underneath the larvae. The rolled cuticle is pushed further ventrad, underneath the larva. When the larva finishes feeding in that crater, it moves to a different spot on the same leaf.

A single larva can spend many days feeding on the same leaf. We observed and filmed the larvae of C. sphaerula cutting and eating trichomes (Video 1). Plants in the tribe Arctotidinae have mostly non-glandular trichomes, although some glandular hairs can be present in certain organs ( Karis et al. 2009). Glandular trichomes would be more deterrent to herbivory. Trichome-eating has not been observed for any Cassidinae . No trichome fragments appear in the shields we dissected (n = 4) so we assume trichomes are digested.

In C. sphaerula , adult feeding resembles larval eating. The adult also makes multiple cuts in the cuticle, in an arc-shape; as it feeds deeper into the trough, the head movements push the cuticle ventrad, under the beetle, towards the posterior margin of that feeding depression (Video 4). The depression deepens and widens, and the cuticle becomes a ridge at the margin of this feeding crater. We did not observe the adults consuming trichomes. The pattern resulting from adult feeding resembles the larval pattern, but the craters are larger. Both stages leave the dorsal cuticle intact, forming windows.

Natural enemies of C. sphaerula . SA observed other animals on the host plant: snails, slugs, spiders, velvet mites, springtails, insects (wasps, aphids, stink bugs, lace bugs, other beetles including one chrysomelid (to be determined), Lepidoptera caterpillars), but noted few interactions that might clarify which are competitors, predators, and parasites of C. sphaerula . We observed and filmed one C. sphaerula larva walking over a leaf and a smallersized aphid moved out of its way (Video 3). In another instance, a smaller mite moved out of the way of an approaching C. sphaerula larva.

Observations of interactions in the field were almost impossible as the host leaves lie flat, pressed against one another and it is necessary to grasp each leaf and gently pull it up to see the underside. This tends to dislodge or scare off many of the other individuals on the plant. The C. sphaerula larvae raise their shields whenever they are disturbed, including by others of the same species. They seem to spend a great deal of time sitting still, but adults are alert—they freeze when there is any movement of the leaf. Then they scuttle to the underside of the leaf, out of the light and view. Like many cassidines, adults show a definite tendency to tumble off the leaf to the ground and then scuttle to the plant stems where they are better protected.

Discussion

The behaviors and life cycle of C. sphaerula was studied in detail and over many months (early spring-late autumn). We confirmed the choice of host plant, A. prostrata , in the indigenous habitat in South Africa; C. sphaerula is now known on two Arcotheca species ( Heron and Borowiec 1997). In South Africa, hosts documented for Cassidini are in the Amaranthaceae , Asteraceae and more infrequent hosts are in Aizoaceae , Fabaceae , Polygonaceae , Salvadoraceae , Solanaceae , and Zygophyllaceae ( Borowiec and Świętojańska 2002 – 2022). This is the second publication to record the feeding habit of C. sphaerula on an Arctotheca species; a comprehensive survey of agents against A. calendula was carried out in South Africa 1986, 1987 and 1988, where C. sphaerula was noted as a potential agent although possibly not sufficiently specific ( Scott and Way 1990). Our study shows a strong association of C. sphaerula with A. prostrata . Further observations and testing of the specificity of C. sphaerula would be necessary to determine whether it could be considered a potential agent for biological control.

Chrysomelid females provide several lines of physical and chemical protection of their eggs, including oothecal and excremental coverings ( Hilker 1994). Eggs have been documented for 13 Cassida species in South Africa and females deposit their eggs singly or in small groups to the undersides of their host leaves (often alongside a vein) in simple oothecae that lacks a stalk. The ootheca of C. sphaerula has a single layer enclosing the eggs, in contrast to the large complex multi-membrane oothecae with many eggs in Conchylotenia ( Heron 1999) and Aspidimorpha ( Muir and Sharp 1904) . In C. sphaerula , ootheca have no fecal cover. Within the genus Cassida , C. coagulata Boheman, 1854 is a notable exception with a larger more elaborate oothecae generally attached to their host plant stem ( Amaranthaceae hosts in this case, not Asteraceae ; H. Heron, pers. observ.). Female oviposition behaviors, including site preparation and coverings of the ootheca, the oothecal structure, and qualities of the egg mass appear to vary within the genus Cassida and suggest novel phylogenetic characters.

We observed the distinct feeding pattern that pushes the epidermis to one side and leaves craters on the dorsal surface of the leaf. Comparison with images and data for other species suggests this is a distinct pattern, now known for at least three South African Cassida species. Author Heron photographed similar patterns for Cassida guttipennis Boheman, 1862 on the host, Berkheya bipinnatifida (Harvey) Roessl (Asteraceae) , and Cassida quatuordecimsignata Spaeth, 1899 on the host, Berkheya maritima J.M. Wood and M.S. Evans (Asteraceae) (see Heron and Borowiec 1997: 643, Fig. 19; Heron 2011: 137, Fig. 9 View Figures 9–15 ; Heron 2003: 43, Fig. XXV) without discussing how the pattern arose. These three species are the only ones where such a pattern is reported; the midrib and secondary veins are not eaten, and the craters are found in areas between veins. Bieńkowski (2010) described the more typical chewing pattern in two other Cassida species. These patterns suggest intrinsic intra-generic variations within Cassida . As more feeding patterns are recognized, novel hypotheses about their significance are emerging; for example, a masquerade strategy in some leaf beetles ( Konstantinov et al. 2018).

The careful observation and filming of feeding in C. sphaerula allow us to determine how the windowpane feeding pattern arises. It is unclear if the rolling over of the epidermis is related to the sheer density of trichomes (see Fig. 13 View Figures 9–15 )—pushing trichomatous cuticle out of the way avoids energy and time costs to cut trichomes and clear a feeding path. We observed C. sphaerula larvae consuming trichomes, which has not been reported for any Chrysomelidae before. In Chrysomelinae chrysomelids, larvae of some Platyphora species were observed to cut and throw trichomes backwards unto their fecal shields ( Bernardi and Scivittaro 1991; Flinte et al. 2017: 15). In Campostomate chrysomelids, larvae trim and store trichomes into a section (“attic”) of the fecal case ( Brown and Funk 2005) or incorporate trichomes and feces to make the case wall ( Chaboo et al. 2008). Trichome-consumption may not be a regular part of the diet and the nutritive value is unclear. The feeding process may be flexible when trichomes are less dense. Trichome density impacts movements of cassidine larval (larvae use the tarsungulus to insert into the epidermis and “tiptoe” to move) ( Medeiros et al. 2004; Medeiros and Moreira 2005). Author Heron’s observations of C. guttipennis feeding revealed that more typical circular feeding scars without rolled cuticle margin is left on those plants with less dense pubescence, e.g., Berkheya speciosa (DC.) O. Hoffm. (Asteraceae) .

In C. sphaerula , all five larval instars and the pupa retain an exuvio-fecal shield. Instar I has a feces-only shield (Fig. 19–20); instars II–V retain previous exuviae in a stack, with feces applied. The pupae exhibit variability, retaining either the instar V exuviae only ( Fig. 26 View Figures 26–31 ) or the entire structure of the larval stages ( Fig. 27 View Figures 26–31 ). It is unclear what the different benefits are of each shield form. Within the genus Cassida , shields vary in architecture, some with exuviae only, or with exuviae covered with fecal or with fecal filaments ( Świętojańska 2009).

Life history data can provide a great deal of comparative information to support species concepts and evolutionary relationships. Some of our findings are relevant to character hypotheses presented in two phylogenetic analyses of Cassidinae , Borowiec (1995), particularly his characters 15–19, and Chaboo (2007; 16 larval characters). Our findings also suggests new characters and new states to expand López-Pérez et al.’s (2018) dataset for the phylogeny of Cassidini . The similarity of feeding pattern in C. guttipennis , C. quatuordecimsignata , and C. sphaerula may be clues to shared behavior and morphology, possibly defining a sub-group within Cassida . The production and relative simplicity of the ootheca in C. sphaerula compared to the more complex one in C. coagulata indicate intra-generic variations and other potential characters, for example ootheca present or absent, size (e.g., number and arrangement of eggs), structure (membranes, additional layers of chewed plant material or feces). The preparation of the oviposition site and the post-ovipositional behaviors of the female await comparative study and evolutionary analysis.

The exuvio-fecal shield that diagnoses the eight derived tribes of Cassidinae is a unique morpho-behavioral complex (Chaboo 2007), an example of an extended phenotype, like a bird’s nest ( Dawkins 1989). This is a significant macroevolutionary event in the evolution of Cassidinae , however, our current picture of its origin is murky. At the base of the tortoise beetle clade, Delocraniini larvae were described as “pouco encobertas pelos excrementos” (=barely covered by excrement) so not carrying a shield ( Bondar 1940: 1 02), Hemisphaerotini larvae have caudal processes and a unique “bird-nest” shield architecture ( Chaboo and Nguyen 2004), and Spilophorini larvae have caudal processes and an exuviae-only shield ( Nishida et al. 2020). In contrast, the mining larvae of Notosacanthini lack caudal processes and lack shields ( Monteith et al. 2021). Also, remarkable is the independent origin of shield retention in the distantly related ‘hispine’, Oediopalpa Baly, 1858 ( Bruch 1906) .

The tortoise beetle shield has been considered as a protection and a defense. Réaumur (1737) hypothesized that it protected against sun and flies. Weise (1893) hypothesized its function as defense against desiccation. More observations led to the hypothesis that shields are a defense against enemies and used cheaply-available defecation products and exuviae and perhaps even chemicals in exocrine glands of those exuviae ( Olmstead 1994). Mechanical defense against predators has been tested experimentally, with support by several researchers ( Eisner et al. 1967; Olmstead and Denno 1993; Eisner and Eisner 2000) but contradicted by others ( Müller and Hilker 1999; Nogueira-de-Sá and Trigo 2002). Further studies with Cassida larvae point to more selective shield defense to certain enemies: Schenk and Bacher (2002) showed shields were effective against vespid predators only, while Bacher and Luder (2005) showed they were effective against parasitoids only and offer some protection against desiccation and wind, but not so against abiotic factors of UV-radiation. Müller (2002) also found variable effectiveness of shields to deter different predators. Chemical defense via enteric discharges in shields was proposed by Pasteels et al. (1988). Chemicals sequestered from host plants or by de novo synthesis can enhance shield defenses ( Gómez et al. 1999; Vencl et al. 1999, 2005, 2009, 2011; Nogueira-de-Sá and Trigo 2002, 2005), however, chemicals were also found to have no impact on larval survival ( Bottcher et al. 2009). This succession of ideas and continuing testing are crucial to illuminating the origin, function (i.e., cost benefit analyses), and diversity of fecal architectures.

Phylogenetic studies in Cassidinae have relied largely on adult characters. In the past, a few characters and states of juvenile stages have been proposed: Borowiec (1995) tested four characters of larvae for his phylogeny of Cassidinae , Chaboo’s (2007) study included 20 from juveniles, López-Pérez et al. (2018) tested one larval character. Going forward, we anticipate more studies like López-Pérez et al. (2021) that hypothesized nine novel characters with their possible states for pupae (their shield present/absent is equal to Chaboo 2007: char. 19). Juvenile stages, behavior, and ecology offer a wealth of new characters that could strengthen systematics of Cassidinae (indeed, all insects), from species concepts to tribal relations. Juvenile stages of most insects are extremely underrepresented in museum collections. The research challenge is detailed field studies and collections and descriptions of specimens.