Ants as shell collectors: notes on land snail shells found around ant nests

Although there is only one myrmecophilous snail species, which has been reported as actually living in ant nests (WITTE et al. 2002), shells of several species are known being gathered by foraging harvester ants. The shell collecting behaviour of ants has been studied by VERDCOURT (1957, 2002), URBAŃSKI (1965), MIENIS (1974) and SEIDL (1987) so far. Messor harvester ants are thought to be typically non-carnivorous species which feed on seeds, fruits and carcasses. Hence, it is puzzling why especially these ants collect shells. They may collect snails because the latter often resemble seeds in size and shape, or they feed on live snails or the carcasses of dead ones URBAŃSKI (1965). We investigated ant nests in semiarid grasslands and shrubs in Turkey in order to better understand the ant–snail relationship in natural habitats. We were interested in gathering information on topics such as whether the samples were biased by the size and/or shape of the snail species. This was done by comparing the snail fauna in or near ant nests with local snail diversity. Furthermore, we speculate on clues to what extent harvester ants consume some snail species.


Introduction
Although there is only one myrmecophilous snail species, which has been reported as actually living in ant nests (WITTE et al. 2002), shells of several species are known being gathered by foraging harvester ants. The shell collecting behaviour of ants has been studied by VERDCOURT (1957VERDCOURT ( , 2002, URBAŃSKI (1965), MIENIS (1974) and SEIDL (1987) so far. Messor harvester ants are thought to be typically non-carnivorous species which feed on seeds, fruits and carcasses. Hence, it is puzzling why especially these ants collect shells. They may collect snails because the latter often resemble seeds in size and shape, or they feed on live snails or the carcasses of dead ones URBAŃSKI (1965). We investigated ant nests in semiarid grasslands and shrubs in Turkey in order to better understand the ant-snail relationship in natural habitats. We were interested in gathering information on topics such as whether the samples were biased by the size and/or shape of the snail species. This was done by comparing the snail fauna in or near ant nests with local snail diversity. Furthermore, we speculate on clues to what extent harvester ants consume some snail species.

Material and Methods
Field sampling was carried out on the area of the Dumlupınar University, Turkey (Fig. 1), between the cami and Hüsnü Özyeğin Öğrenci Yurdu (student dormitory), 10-15 October, 2007 on an altitude of about 1050 m above Malacologica Bohemoslovaca (2009), 8: 14-18 ISSN 1336 sea level (geographic coordinates: 39. 4811°N,29.8889°E (using Google Earth). The habitat of about five hectares, which consisted of a relatively uniform secondary steppe with rocks and shrubs (mainly Juniperus, Quercus, Crataegus) in some places. Ant nests were evenly distributed over the area, with an average distance of 10-14 m from each other. Entrance mounds of the largest nests could be 1 m in diameter. In general, ant nests generally consisted of a circa 0.3 m high central mound surrounded by a small amount of litter. This litter contained remnants of seeds and fruits (mostly Poaceae, Asteraceae, Medicago, Rumex, and Triburus). Litter composition around nests may vary greatly, because different colonies collect different seeds (TRANIELLO & BESHERS 1991). The demarcation of a single nest-mound was sometimes difficult due to the dispersed nest entrances (HELLER 1971). Therefore it was also difficult to find the center of the nests. However it was usually marked by well visible foraging routes around the nest entrance. We used 10 paired quadrats (0.25×0.25 m each) to collect litter from the soil surface. Each quadrate pair consisted of a near (max. 0.3 m from the nest entrances) and a far (3 m from the nest entrance) samples. In this way we sampled 10 ant nests. We then compared the samples to see which snail species were collected by the ants and whether ants collect shells selectively from the available overall local snail population. Besides the quadrate pairs, we investigated shells found around nests in three habitat types: shrubs, rocky steppe and steppe. We collected 3 × 1 litter samples from each habitat type. In order to have more data on ant nests, we collected further 47 × 1 litter samples from the vicinity of ant nests in the steppe habitat (altogether 50 samples with the previous three). We complemented quadrate samples by visual search in the field to get a more complete list of the snail fauna. All the shells were sorted and identified in the laboratory. We could not find living specimen of Monacha (Paratheba) bithynica Hausdorf, 2000, possibly because of the late autumn season, when the species is not active. The identification of the species was dubious, because it is hard to distinguish between this and its congeners (M. margarita, Hausdorf, 2000, M. crenophila (L. Pfeiffer, 1857), M. ovularis (Bourguignat, 1855)) based solely on shell characters. However, HAUSDORF (2000) found this species in nearby areas. One living specimen of Monacha (Monacha) solidior (Mousson, 1863) was found, thus its identification is supported by anatomical evidence. Identity of juveniles of Helix lucorum Linnaeus, 1758 was based on the two adult shells found in the territory. We only counted the number of shell apices to avoid multiple counting of the same shell. For the nomenclature of the snail species we used the work of SCHÜTT (2001). Collected Messor specimens are deposited in the Hymenoptera collection of the Hungarian Natural History Museum, and the snail shells are in the private collection of the senior author.

Results
We collected ant workers and soldiers which were stored in alcohol. We found two ant species Messor oertzeni Forel, 1910 andMessor caducus (Victor, 1839) in the area, but their relative abundances was not determined. From several nests 20 specimens of M. oertzeni and 41 specimens of M. caducus were collected. Empty shells were observed on anthills of both species. Samples from the three habitat types revealed nine snail species (Table 1). The number of species in the grassland and rocky grassland habitats was low (four and three species, respectively) compared to the shrubs (eight species). Zebrina kindermanni (L. Pfeiffer, 1850) was found only  in the grassland, Chondrus zebrula tantalus (L. Pfeiffer, 1868) was most common in the rocky steppe. Xerolenta obvia, M. bithynica, and Ch. zebrula tantalus were equally common in the steppe habitat. Besides the grassland species, five species were found only in the shrubs (Table 1). The 50 soil samples collected in the steppe habitat revealed seven species. The most common species were the same as in the above comparison: X. obvia, M. bithynica, and Ch. zebrula tantalus. In addition four species occurred infrequently (Fig. 2). A total of 50 Monacha bithynica, eight Xerolenta obvia (Menke, 1828), two Helix lucorum and one Helicopsis subcalcarata neuberti Hausdorf, 1990 shells were found in the ten near samples of the paired quadrates, while we found only one M. bithynica in the ten samples three meters apart from the nests. During a complimentary search, we found the five shrub species in the steppe areas as well except for Oxychilus (Mediterranea) hydatinus (Rossmässler, 1838). These were however eroded shells found around small shrubs. In addition we also found some specimens of Zebrina detrita (O.F. Müller, 1774), H. subcalcarata neuberti, and Helix (Pelasga) escherichi O. Boettger, 1898 during the visual search, which were not encountered in the litter samples. The species Zebrina detrita was found only with additional search besides bushes. The species is probably too big to be transported by ants. One specimen of Zebrina kindermanni was found in a litter sample from the grassland. It was rarely found during visual search. The thin shelled species Monacha bithynica was the most common species in all habitat types. Sixty percent (173 out of 310) of the shells were perforated in a similar way, on the penultimate whorl (Fig. 3). The species M. solidior was found only in the shrubs. The shells were perforated as in the other Monacha species. It is likely that the perforations are made by ants to reach the soft body of the snail.
Chondrus zebrula tantalus was common in ant nest garbage. Most of the specimens were juveniles with few whorls. We found only 73 fully grown undamaged adults out of 265 individuals. Apertures of the juvenile shells were filled by mud, thus presumably these were collected by ants in a not so fresh stage, and probably the shells were found in the soil during digging the ducts. The shells of the rare species Multidentula ovularis (Olivier, 1801) were clean and undamaged. One specimen of Oxychilus hydatinus was found during visual search from nest garbage. The species lives underground and is rarely found alive (KERNEY et al. 1983). In the shrubs the species was more common and was also found  in nest garbage. Most probably the ants found these shells in the underground ducts. The species Helicopsis subcalcarata neuberti was rarely found only during the visual search. It is more abundant in the shrubs. Shell apertures were dusty, thus ants might find them during nest construction. Xerolenta obvia was common; almost all the specimens possessed only few whorls, and were less than 0.7 cm in diameter. Helix escherichi was found only during the visual search, especially in the shrub habitat. During this search, two adults were found of the species H. lucorum. Freshly hatched juvenile (embryonic) shells were abundant in nests.

Discussion
As a result of different shell collecting activities of ants, all of the shells can be found around the nests, together with unconsumed vegetable parts. Some shells have been collected distant from the nests, whereas others might be found during building the nest under the soil. The determination of the origin of the collected shells is not obvious, so we can just conjecture it knowing the lifestyle of the snail species.
The shell collecting activity of ants is not unique for the investigated area. VERDCOURT (1957VERDCOURT ( , 2002  The previous results match with our observations. (1) The fauna of the area consists of more species that were found around the ant nests.
(2) Large bodied species were found less frequently in the nests.
(3) Most shells resembled seeds. But we found no clue about the significance of snails in the food of ants. We conclude that shells found on the area of ant nests represent only part of the local snail fauna, viz. the smaller species or juveniles from larger species. Further, it seems plausible to state that some snail species with thin shells may be consumed regularly by ants. Thus, when dealing with shells found around the ant nests (which is appealing because of the high concentration of shell compared to average density in the surroundings), we should take into account the shell selectivity of ants. From a fauna inventory perspective, ants may help malacologists to find species that live underground.