Mantodea (praying mantises) is a small but prominent insect order comprising approximately 2,400 species worldwide (Wieland and Schütte 2012, Wieland 2013). From South Africa, 157 mantis species belonging to 64 genera in 14 families have been recorded (Greyvenstein et al. 2023). Three of the 11 species of the genus Polyspilota Burmeister, 1838 are known from South Africa (Greyvenstein et al. 2023, Otte et al. 2025). These [sub-]species are Polyspilota aeruginosa aeruginosa (Goeze, 1778), Polyspilota caffra caffra Westwood, 1889 and Polyspilota magna Giglio-Tos, 1911

(Greyvenstein et al. 2023). The latter two species are suspected to be endemic to southern Africa (Kaltenbach 1996), whereas P. aeruginosa occurs in many sub-Saharan countries (Ehrmann 2002). Polyspilota magna is known only from the female holotype (type locality ‘Cape of Good Hope,’ South Africa) (Natural History Museum, Berlin, Germany, ZMHB) (Giglio-Tos 1911, 1927, Otte et al. 2025). Despite their conspicuousness, information on the distribution, ecology, behavior, and feeding strategies of Polyspilota species remains limited.

Mantises are primarily sit-and-wait predators that rely on crypsis and specialized raptorial forelimbs to capture prey. This, by default, makes these insects opportunistic and generalist hunters of smaller invertebrates. A wide range of unusual prey species has been reported for Mantodea. Several reports indicate that some mantids may also feed on small frogs (Nash 1962, Ridpath 1977, Costa-Pereira et al. 2010), lizards (Jehle et al. 1996, Mebs et al. 2016), small birds (Ridpath 1977, Nyffeler et al. 2017), mice (Nickle and Harper 1981, Ehrmann and Schmidt 1992), snakes (Kevan 1985), and even fish (Battiston et al. 2018). A mantis species often reported to feed on many kinds of prey and on unusual prey is the Afrotropical Polyspilota aeruginosa (Goeze, 1778). This mantid has, among many others, been observed to feed on a juvenile songbird (Beier 1969, Nyffeler et al. 2017) and to exhibit startle-threat behavior to deter the attack of a dark-capped bulbul Pycnonotus barbatus (Desfontaines, 1789) (Grandcolas and Desutter-Grandcolas 1998).

In this paper, we report unusual feeding behavior by Polyspilota species preying on snails and slugs (Mollusca: Stylommatophora) in South Africa. To our knowledge, this is the first published record—globally—of snail eating among Mantodea in the wild. We also expand on this behavior as previously recorded in captivity (Reitze and Nentwig 1991). Beyond documenting predation on snails, we compiled incidental records of the feeding behavior of South African Polyspilota species to assess whether they exhibit broader opportunistic predation on atypical prey. Discovering novel feeding strategies is of biological interest as it would contribute to better understanding of the theoretical niche a species may occupy (Raes et al. 2008). Therefore, this study aims to help bridge this knowledge gap by providing insights into the distribution, behavior, and feeding strategies of Polyspilota in South Africa. Observations of mantids preying on snails also contribute to our broader understanding of the opportunistic feeding strategies of mantids.

A single mantid nymph was observed at the iGwalagwala Trail (St Lucia, KwaZulu-Natal, South Africa) feeding on a snail on 10 March 2022. Both the mantid and the snail were collected and accessioned into the KwaZulu-Natal Museum (accession numbers NMSA-Man 000475 and NMSA-Mol 0P2042, respectively). We also reared a Polyspilota nymph in captivity and provided snails as prey (Fig. 4). The aforementioned snails were identified by Dr. I.V. Muratov of the KwaZulu-Natal Museum, Pietermaritzburg, South Africa, using descriptions and the keys presented by Herbert and Kilburn (2004).

Distribution data of Polyspilota species in South Africa were obtained from specimen collections and the internet platforms Global Biodiversity Information Facility (GBIF) and iNaturalist. The distributions of the South African Polyspilota species were mapped with ArcGIS Pro by Esri (2025 California, USA, https://www.esri.com/en-us/home). Records of Polyspilota oothecae on iNaturalist were disregarded since Polyspilota oothecae cannot be reliably distinguished from those of other mantid species based only on external morphology. For iNaturalist records, additional information was recorded where available, including habitat type, whether the mantid was observed feeding, and what the prey was. Prey species were identified to the lowest possible taxonomic level (order). Unusual or infrequent prey species were included to provide a broader context for the trophic ecology and feeding strategies of the Polyspilota species. Museum records of Polyspilota specimens were taken from Greyvenstein et al. (2023), who compiled data from the following South African institutions: Ditsong National Museum of Natural History, Pretoria (TMSA), South African National Collection of Insects, Pretoria (SANC), National Museum, Bloemfontein (BMSA), Albany Museum, Makhanda (AMGS), Rhodes University, Department of Zoology and Entomology, Makhanda (RUDZ), Durban Natural Science Museum, eThekwini (DMSA), Iziko South African Museum, Cape Town (SAMC), and KwaZulu-Natal Museum, Pietermaritzburg (NMSA).

The most comprehensive taxonomic key to species of Polyspilota remains that provided by Giglio-Tos in his revision of Mantidae (Giglio-Tos 1927), which largely reiterates the species concepts and diagnostic characters presented in Giglio-Tos (1911). Some additional specimens and measurements were added for P. aeruginosa in the revision (Giglio-Tos 1927), while no additional information was added for P. caffra or P. magna. We used these keys to identify the South African Polyspilota individuals available on iNaturalist. The principal diagnostic feature discriminating these three Polyspilota species is the relative length of the pronotal metazone (Giglio-Tos 1911, 1927), as expressed in the ratio of the metazone to the prozone (Fig. 1, Table 1).

Additionally, P. aeruginosa and P. caffra have a distinct black spot on the ventral surface of the profemur between the profemoral base and the ungual groove, which is absent in P. magna. Since we did not encounter any mantis individuals corresponding morphologically to P. magna, we present only the diagnostic ratios for P. aeruginosa and P. caffra (Table 1). We identified to species level only those iNaturalist records in which individuals could be viewed dorsally and measured reliably. Other records were retained at the genus level. The metazone:prozone ratios clearly separated the two species involved without overlapping values among sexes. The ratio between pronotum length and pronotal width provided additional discrimination (Table 1).

In total, 466 records of Polyspilota spp. from South Africa, Eswatini, and Lesotho were available for analysis. Grouping these countries into a unit is standard practice in conservation assessments (Lee et al. 2025). Most of the records (351) were retrieved from iNaturalist, whereas 101 records originated from museum collections in South Africa and only 14 records were uniquely obtained from the GBIF (2025). The GBIF data included seven specimens from museum collections outside South Africa and seven DNA sequences recorded in the BOLD (Ratnasingham and Hebert 2007, Ratnasingham et al. 2024). The mantises in 140 iNaturalist observations could be identified as P. aeruginosa and 73 as P. caffra, but the mantises in an additional 138 observations could be confirmed only to the level of genus, mostly because the angle of the corresponding citizen scientist photos precluded the measurement of the diagnostic ratios of the pronotum.

Distributions ofPolyspilotaspecies.—The distributions of the Polyspilota species in South Africa include the eastern seaboard and low-lying subtropical regions east of the eastern Great Escarpment, with only five records from the country’s northeastern interior (Gauteng) (Fig. 2). Polyspilota aeruginosa and P. caffra are largely allopatric. Polyspilota aeruginosa mainly occurs in the northeastern coastal region of KwaZulu-Natal and the northern parts of the Limpopo and Mpumalanga provinces, whereas P. caffra has been recorded from the southeastern and southern coastal regions of the Eastern Cape and Western Cape provinces (Fig. 2). Polyspilota aeruginosa is most abundant in KwaZulu-Natal (70% of the records), followed by Limpopo and Mpumalanga (19%), the Western and Eastern Cape (6%), Eswatini (5%), and Gauteng (<1%). In contrast, P. caffra mostly occurs in the Western and Eastern Cape (87%), followed by Limpopo and Mpumalanga (9%) and KwaZulu-Natal (4%). Polyspilota caffra has not yet been recorded from Gauteng or Eswatini. The distributions of the South African species of Polyspilota suggest that this genus is eurytopic, also occurring in suburban gardens, urban areas with semi-natural vegetation patches, vegetation patches between agricultural fields, and natural areas (Table 2).

Most records available to us (36%) were from urban areas with patches of semi-natural vegetation. Polyspilota caffra was observed more frequently (20%) in botanical gardens and national parks, while P. aeruginosa was more frequently recorded in urban areas with semi-natural vegetation patches (Table 2). In the entire dataset, the natural areas were predominantly subtropical coastal habitats, whereas the corresponding records from Mpumalanga and Limpopo were typical savanna habitats. All suburban gardens with observations were characterized by abundant trees and shrubs.

Polyspilotaspecies preying on mollusks.—We believe that this study is the first report of any Mantodea feeding on any mollusks in the wild. The observation made under entirely natural circumstances was of a P. aeruginosa nymph feeding on a silky tailwagger snail Kerkophorus inunctus (Melvill & Ponsonby, 1899) (Stylommatophora: Urocyclidae) in foliage at St. Lucia, KwaZulu-Natal, South Africa (Fig. 3). The photographs provided in Fig. 4 are those of a P. aeruginosa nymph reared in captivity. This nymph readily accepted common garden snails Cornu aspersum (Müller, 1774) (Stylommatophora: Helicidae) as food. The mantid grabbed the snail before it fully retracted into its shell, firmly securing it with its raptorial forelegs and maintaining constant pressure to prevent escape. Feeding was focused on the exposed tissue, and once this was consumed, the mantid discarded the snail and its shell.

Of the almost 1,200 iNaturalist observations of Polyspilota mantises in South Africa as of February 2026, 17 showed mantises eating “typical” kinds of prey, whereas five showed them to feed on snails or slugs (Table 2). These observations included three P. aeruginosa individuals, of which two were nymphs and one was an adult male. Two of these observations were made in KwaZulu-Natal (Ballito and Scottburgh) and one in Limpopo (Hoedspruit). The remaining two observations were of a male P. caffra in George (Western Cape) and one of an undetermined species of Polyspilota nymph in Kelso (KwaZulu-Natal) (Table 2).

In addition to observations of mollusk-feeding mantises, several records of Polyspilota species consuming other invertebrates have been recorded (Table 2). We received photos of a female P. aeruginosa capturing and consuming a tropical house gecko Hemidactylus mabouia (Moreau de Jonnès, 1818) on 14 May 2025 in Leisure Bay, KwaZulu-Natal (Fig. 5). Similarly, four days prior to the latter observation, a Polyspilota cf. caffra female was recorded catching, dismembering, and eating a marbled leaf-toed gecko Afrogecko porphyreus (Daudin, 1802) (Table 2). Although predation on small vertebrates has previously been reported for several mantid species, this appears to be the first documented case of reptile predation by P. aeruginosa in South Africa.

Although three Polyspilota species are suggested to occur in South Africa (Greyvenstein et al. 2023), none of the museum specimens or iNaturalist records examined in this study matched the diagnostic description of P. magna (Giglio-Tos 1911, 1927). It remains unclear whether this is, indeed, a valid species. It was described from a single specimen with similar proportions to those of P. caffra, the only notable difference being the absence of a black spot on the ventral surface of the profemur, anterior of the trochanter. Giglio-Tos (1911) subsequently indicated that this may have been a simple anomaly and that more individuals need to be examined to clarify the identity of this taxon.

This study confirms the distributions of the three South Africa Polyspilota species as furnished by Beier (1935) and Ehrmann (2002). In South Africa, the coastal distribution of this genus indicates that it occurs predominantly in eastern, subtropical regions and in the eastern savanna biome in the northern parts of the country. It is noteworthy that the apparent separation in the distributions of P. aeruginosa and P. caffra corresponds closely to the border between KwaZulu-Natal and the Eastern Cape province, which follows the Umtamvuna River gorge. This gorge is recognized as an important biogeographical barrier that limits gene flow in a range of taxa (Clancey 1994, Suinyuy et al. 2012). Vegetation differences across this boundary may additionally explain species distributions within the genus Polyspilota. KwaZulu-Natal supports more extensive and continuous forests, including scarp, mistbelt, and coastal-dune forests, due to higher rainfall, whereas the Eastern Cape contains smaller, fragmented forest patches embedded in a landscape of grassland and thicket (e.g., Parker et al. 2022). Most P. caffra records are from the Western Cape province, which is dominated by the Fynbos biome.

This study is the first to report the novel snail-eating behavior of Polyspilota mantises in the wild as well as confirmation of this behavior in captive mantids. Laboratory studies, which included six Mantodea species, have suggested that P. aeruginosa has an unusually broad diet and is the most generalist feeder (Reitze and Nentwig 1991). This species ate 75% of all the prey species it was offered in captivity, including two slug and one snail species. The other five mantid genera had an acceptance rate of between 43% and 69% across all prey items, and a particular low acceptance rate of snails and slugs which the authors attributed to their slow movement (Reitze and Nentwig 1991). The field observations reported here complement these experimental findings by demonstrating that Polyspilota mantises freely exploit gastropod prey under natural field conditions.

The sit-and-wait predatory strategy of Mantodea relies on striking prey in response to visual cues, particularly prey movement (Gelperin 1968). Mantises attempt to catch prey when several visual stimulus parameters are met (Prete et al. 2002, 2011). These include the overall size of prey, its direction of movement, and its apparent speed. Slower-moving targets generally elicit fewer strikes (Prete et al. 2011). Hence, it is unusual for slow-moving prey, such as slugs and snails, to trigger mantises into action. However, the strike speed and structure of P. aeruginosa have been reported to vary depending on the movement of the prey, with slower prey associated with slower tibial extension and more frequent pauses (Rossoni and Niven 2020). This behavioral flexibility allows timing adjustments of the strike, particularly the final downward sweep of the tibia (Rossoni and Niven 2020). This behavioral adaptation, which has only been reported in three other mantid species, could allow members of the Polyspilota genus to capture slow-moving prey such as snails and slugs.

It remains unclear whether snail feeding is opportunistic or whether it represents a more common, hitherto overlooked, component of the diet of these mantises. This behavior could also be a response to unfavorable conditions and the associated scarcity of prey. Terrestrial snails are generally abundant in moist environments (Chiba 2007), which corresponds well with the distribution of Polyspilota species in South Africa. These observations highlight the opportunism of mantids also reported by Reitze and Nentwig (1991), suggesting that their diet composition may be strongly influenced by local resource availability and climatic conditions that promote gastropod abundance.

The catching and eating of snails by Polyspilota mantises based on observations in South Africa suggest that this remains the only genus observed to do so in the wild. It was suggested by Henríquez-Piskulich et al. (2023) that feeding on snails might be uncommon among mantises due to the mucus coating on the bodies and shells of snails, reflecting light. Most studied mantises have been reported to be less likely to track and strike at glossy targets (Henríquez-Piskulich et al. 2023) and slow-moving prey (Gelperin 1968, Reitze and Nentwig 1991, Prete et al. 2011). These factors do not seem to deter Polyspilota mantises from catching and eating snails, but the mechanisms for this remain unknown. Opportunism or instinct may have facilitated the development of this novel feeding strategy, as highlighted by the observation of a female P. aeruginosa catching and eating an entire gecko (Fig. 5).

The nutritional content of snails generally indicates that they contain much more than mucus, moisture, and fat. Rygało-Galewska et al. (2022) suggested that the common garden snail (Cornu aspersum), on average, consists of 78–83 g of water, 11–17 g of protein, and only 0.4–1.2 g of fat per 100 g. The protein of C. aspersum was also indicated to be highly digestible, and 100 g of this snail tissue contains approximately 30% of the daily dietary essential amino acid requirements for humans (Surówka et al. 2025). Snails also contain minerals such as calcium, sodium, potassium, magnesium, and iron (Nkansah et al. 2021, Pathak et al. 2025).

There are, however, risks to mantises eating snails. For example, snails are known to be the intermediate hosts of various parasitic nematodes and trematodes (Lu et al. 2018). Although these parasites are known to infest other terrestrial insects, limited information exists about parasites of mantids. The only parasite known to infest P. aeruginosa is a nematode horsehair worm, Chordodes ferox Camerano, 1897 (Nematomorpha: Gordiida) (Schmidt-Rhaesa and Perissinotto 2016). The unusual snail-eating behavior of Polyspilota reported in this study generates new avenues for future research into mantid feeding strategies and the possible risks mantids could be exposed to due to their interactions with terrestrial gastropods and their related parasites.

These observations confirm that Polyspilota mantids are opportunistic feeders capable of exploiting novel locally available food resources. Feeding on mollusks holds the benefit of high-protein and mineral-rich prey, but it also increases the risk of parasite infestation. Any such potential trade-off remains to be explored. Knowledge about the ecology and behavior of Polyspilota species is limited, as is knowledge of most South African Mantodea and their ecobiology and lifestyles.