Journal of Insect Behavior

, 20:367

Trail-Following Behaviour in the Malacophagous Larvae of the Aquatic Sciomyzid Flies Sepedon spinipes spinipes and Dictya montana


    • Department of EntomologyUniversity of California
    • Applied Ecology Unit, Centre for Environmental ScienceNational University of Ireland
  • Timothy D. Paine
    • Department of EntomologyUniversity of California
  • Michael J. Gormally
    • Applied Ecology Unit, Centre for Environmental ScienceNational University of Ireland

DOI: 10.1007/s10905-007-9083-2

Cite this article as:
Mc Donnell, R.J., Paine, T.D. & Gormally, M.J. J Insect Behav (2007) 20: 367. doi:10.1007/s10905-007-9083-2

The ability of neonate larvae of the aquatic sciomyzids, Sepedon spinipes spinipes (Scopoli) and Dictya montana Steyskal (Diptera), to follow snail mucus trails was assessed using filter paper Y-mazes. On finding a mucus trail, larval behaviour of both species switched from unstimulated to stimulated searching behaviour, the latter being characterised by an increase in larval velocity and the frequency of lateral head taps. When fresh mucus trails were used, all of the neonates displayed a positive response and followed the mucus trail into the experimental arm. In addition, for S. s. spinipes and D. montana 80.00% and 86.67% of larvae respectively exhibited a strong response and followed the trail to its end. The stimulatory substance (s), however, appears to become inactive with time and after 45 minutes none of the tested larvae reached the trail end. These results are discussed in relation to the ability of aquatic species to forage outside of water for prey and the implications for their use in the biological control of nuisance snails.


DipteraSciomyzidaemalacophagysnail mucustrail-following behaviourbiocontrol


Trail-following behaviour is a relatively well known behavioural trait in insects. It has been documented for a wide range of taxa including immature Lepidoptera (Ruf et al., 2001), cockroaches (Blattodea: Miller and Koeiller, 2000), myrmecophilous ground beetles (Coleoptera, Carabidae: Cammaerts et al.,1990), rove beetles (Coleoptera, Staphylinidae: Quinet and Pasteels, 1995), ants (Hymenoptera, Formicidae: Hölldobler and Wilson, 1990) and termites (Isoptera: Ohmura et al., 1995). The ability to follow gastropod mucus trails has also received much attention. Such behaviour is used by the pulmonate slug (Mollusca, Gastropoda), Limax pseudoflavus Evans (=Limax maculatus (Kaleniczenko)), and is thought to play a major role in courtship, food location and in homing (Cook, 1992). It is also utilised by arthropod snail-killing predators such as the Lampyridae (Coleoptera: Schwalb, 1961).

Adults of the family Sciomyzidae (Diptera) are found in a wide variety of moist to aquatic habitats where hygrophilous snails occur (Rozkošný, 1999). Females of many species oviposit onto emergent vegetation, and neonates crawl downwards to the water surface. The larvae are relatively unique amongst insects in that they are almost exclusively malacophagous (Knutson and Vala, 2002). The vast majority feed as aquatic predators or terrestrial parasitoids (and/or predators) on snails and a few attack slugs, fingernail clams (Sphaeriidae) or feed on snail eggs. The aquatic snail-killing species vigorously attack prey within the families Lymnaeidae, Physidae and Planorbidae generally at the water surface. Most are generalist predators and have been known to attack and feed on stylommatophorans in addition to their usual basommatophoran prey. In the laboratory, Lindsay (1982) and Beaver (1972) recorded Ilione albiseta (Scopoli) as feeding on the succineids Oxyloma elegans (Risso) and Succinea putris (L.) respectively. Berg (1953) showed that Sepedon fuscipennis Loew could feed on Oxyloma decampi Tryon and Neff and Berg (1966) cited Oxyloma retusa (Lea) and the agriolimacid slug, Deroceras laeve (Müller) as suitable prey for Sepedomerus macropus (Walker).

Prey location by aquatic sciomyzids was investigated by Appleton et al.(1993) who described both passive and active behaviours. Sepedon scapularis Adams actively pursued its prey in the water, whereas Sepedon neavei Steyskal suspended itself from the water surface by means of its hydrofuge interspiracular processes and ambushed snails passing below. In addition, it was recently postulated by Knutson and Vala (in press) that the larvae of aquatic species and their snail prey are brought together at the base of emergent vegetation by the surface tension of the water. This phenomenon, which also occurs in mosquito larvae (Hess and Hall, 1943), would greatly facilitate snail location by sciomyzid species. However, Mc Donnell et al.(2005a) suggested that there is likely to be extensive overlap between the microhabitats of aquatic and semi-aquatic sciomyzids. Such a hypothesis is strengthened by Lindsay (1982) and Barraclough (1983) who found larvae of the aquatic Ilione albiseta (Scopoli) and S. neavei respectively, foraging for food out of the water on wetland vegetation in the wild.

Since sciomyzid larvae feed almost exclusively on molluscs, mucus trail-following may also seem like a logical adaptation for snail location. Coupland (1996) investigated the effects of snail mucus and faeces on both larval behaviour and adult oviposition of the terrestrial sciomyzid, Pherbellia cinerella Fallén. He found that the mucus and faeces of the snail Cernuella virgata (da Costa), stimulated increased searching behaviour in first instar larvae and adult females oviposited more frequently on substrates containing fresh snail mucus. Trelka and Berg (1977) investigated the response of rehearsed (larvae having slug contact shortly before the test) and unrehearsed larvae of the slug killing sciomyzid, Tetanocera plebeja Loew to slug mucus trails. Two trail types were used, normal and supernormal (a second trail painted on top of a normal trail). Larvae were either placed directly onto the slug trail or made to approach it from an angle of 90°. The authors concluded that third instar T. plebeja larvae followed fresh mucus trails only after recent contact with a slug and they often followed in the wrong direction.

To date, the ability of aquatic and semi-aquatic sciomyzid larvae to utilise snail mucus trails for prey location has not been examined. The aim of this paper is to investigate the potential of neonate larvae of the aquatic species, Dictya montana Steyskal (Nearctic) and Sepedon spinipes spinipes (Scopoli) (Palaearctic), to follow fresh snail mucus trails out of the water. The impact of trail age on larval trail-following behaviour is also considered.


Sciomyzid Collection and Rearing

Dictya montana adults were collected using a sweep net on the emergent vegetation of a drainage ditch at the Agricultural Operations Station (AgOps), University of California, Riverside during October 2001. In the laboratory, the flies were kept in a glass beaker containing short pieces of wood which acted as resting sites, two pieces of cotton wool (one soaked in deionised water and the other smeared with a honey/yeast food mixture in a 3:1 ratio) and a freshly crushed Lymnaea snail. The beaker was covered with a gauze lid. The water-soaked cotton wool was replaced every second day and the honey/yeast mixture and crushed snail were replaced twice weekly. In the case of S. s. spinipes, this species was being cultured in the laboratory at the National University of Ireland, Galway so adults were readily available and were maintained in glass jars as described above. Jars of both species were checked daily for eggs which were transferred with a moist paintbrush to a Petri dish containing deionised water. The Petri dishes were then submerged in weighted plastic containers in a waterbath (26°C). These were checked daily and neonates that hatched were used in trail-following experiments. In California, Lymnaea sp. (the taxonomy of North American stagnicoline lymnaeids is unresolved because of their poorly understood morphological variabiliy (Barnes, 1990)) was also collected at AgOps and was used as the test snail, while in Ireland the test species was Lymnaea stagnalis (L.). The latter was collected in a small pond close to the university. Both species were kept in coldrooms in buckets containing water and vegetation from the collecting sites.

Assay Protocols

Neonate trail-following behaviour was assessed using filter paper Y-mazes (Fig. 1) that were saturated with deionised water. Mazes consisted of a single stem and two arms, each of which was 3 cm long, 1 cm wide and separated by an angle of 30°. A moist paintbrush (Type AF85) was used to remove mucus from the foot of three snails (8 cm long approx.) and a mucus trail was then painted onto one of the arms (chosen at random) of the Y-maze. These sizes and numbers of snails were necessary to ensure a complete experimental trail. The trail extended the length of the arm and across the top of the stem thereby ensuring that the larva came into contact with it (Fig. 1). Each set of three snails was used to paint only one trail and a new set was used for each additional trail. The other Y-maze arm had no mucus trail and acted as a control. In investigating the impact of trail age on neonate trail-following behaviour, additional mucus trails were left for 45 minutes prior to testing. For both fresh and aged trails, larvae were accustomed to the test environment by placing them on a saturated (deionised water) filter paper disk in a Petri dish for 15 minutes before testing. Neonates were then placed at the base of the Y-maze stem and encouraged forward using a stationary photo-optic light source (sciomyzid larvae are negatively phototactic).
Fig. 1.

Filter paper Y-maze.

If the larva followed the trail into the experimental arm, it was deemed a positive response and a non-response occurred when the larva crawled over the trail, did not follow it and moved randomly over the control arm or onto the Petri dish arena. A positive response was further subdivided into:
  1. (a)

    Strong response: neonate followed the snail mucus trail to its end.

  2. (b)

    Weak response: larva followed the mucus trail but deviated before reaching the end. On leaving the trail the larvae tended to crawl randomly away from it and often moved completely off the Y-maze on to the Petri dish test arena.

Larvae were inexperienced, were tested individually and each larva was tested only once. After each assay, the Y-maze and test arena were replaced to prevent cross contamination and pseudoreplication. For each test, the experimental and control arms were selected at random to account for any larval side preferences. Fifteen larvae, each representing an individual replicate were tested against fresh mucus trails and another fifteen were tested against aged mucus trails. The data were analysed using a chi-squared test and the standard levels of significance used throughout were P < 0.05, P < 0.01 and P < 0.001.
Table I.

The Number of Neonate Larvae of Sepedon spinipes spinipes (Scopoli) and Dictya montana Steyskal Which Exhibit Positive and Non-Responses to Fresh and Aged (45 minutes) Snail Mucus Trails. Positive Response Further Divided into Strong (Followed Trail to End) and Weak (Deviated from Trail Before Reaching End) Responses


Positive response


Fresh Trail

Sepedon s. spinipes



  Strong response

80.00% (12)

  Weak response

20.00% (3)

Dictya montana



  Strong response

86.67% (13)

  Weak response

13.33% (2)

Aged Trail

Sepedon s. spinipes



  Strong response


  Weak response

93.33% (14)

Dictya montana



  Strong response


  Weak response

86.67% (13)

Note: Chi-squared analysis incorporating Yates correction factor for continuity. Values with the same superscript letter indicate a significant difference between the larval responses.

Fresh trail: Sepedon spinipes spinipesa: χ2 = 13.06, 1df, P < 0.001; Dictya montanab: χ2 = 13.06, 1df, P < 0.001.

Aged trail: Sepedon spinipes spinipesc: χ2 = 9.60, 1df, P < 0.01; Dictya montanad: χ2 = 6.67, 1df, P < 0.01.


The behaviour of S. s. spinipes and D. montana neonates prior to coming into contact with a mucus trail can be defined as unstimulated searching behaviour. Larvae moved relatively slowly and the number of lateral head movements (head tapping) was moderate. On discovering a fresh mucus trail, larvae of both species ceased moving momentarily and appeared to investigate the trail with their mouthparts. This was defined as a trail detection response. Neonates then changed their direction of motion to follow the trail and increased both their velocity and rate of head tapping. This was termed stimulated searching behaviour. These results concur with Coupland (1996) who found that the number of P. cinerella larval head taps was significantly greater in the presence of both snail (C. virgata) mucus and faeces than with deionised water. In this study, all of the tested larvae displayed a positive response (i.e. followed the fresh mucus trail) with 80.00% and 86.67% of S. s. spinipes and D. montana larvae respectively following the trail to its end (Table I). In fact, significantly more larvae of both species exhibited a positive response (S. s. spinipes: χ2 = 13.06; D. montana χ2 = 13.06: 1df, P < 0.001) than a non-response. While following the fresh mucus trails, the vast majority of neonates displayed behavioural patterns that were common to both species. Initial trail-following tended to be vigorous with larvae often circling or doubling back within the first 15 mm of the trail. Similar results were also reported by Trelka and Berg (1977) for third instar larvae of T. plebeja. This behaviour at first may seem unproductive and both time and energy consuming but it may be attributable to the fresh mucus used to mark the trail. In nature, such fresh mucus would indicate that the snail is still in the immediate environment and a thorough search of the trail initially is likely to yield snail prey. Alternatively, such behaviour may be used by neonates to determine the direction of motion of the snail which would be essential for successful prey location. Larvae of both species also appeared to be able to detect the physical boundaries of the fresh mucus trail. Neonates frequently breached the trail edge which resulted in a rapid, instantaneous retraction of the larval head and anterior end followed by a change of direction back onto the trail. In addition, on reaching the trail end, neonates ceased moving, extended their anterior end forward and moved it in an arc investigating the surrounding environment. Larvae, however, did not caste (lifting of the anterior end and striking out forwardly or laterally with their posterior segments fixed to the substrate) as highlighted for T. plebeja by Trelka and Berg (1977).

Larvae of S. s. spinipes (Neff and Berg, 1966) and D. montana (Mc Donnell et al., in press) are known aquatic predators but the ability of these species to follow fresh snail mucus trails out of the water (in this case on filter paper) suggests that they may have the potential to forage for food in shoreline and semi-aquatic areas. Such foraging behaviour in the wild has also been reported for other aquatic sciomyzids by Lindsay (1982) and Barraclough (1983) for I. albiseta and S. neavei respectively. Since sciomyzid larvae are almost exclusively malacophagous, they may have the potential to be used as biological control agents of snail hosts of trematode diseases (Mc Donnell et al., 2005b; Barker et al.,2004; Gormally, 1985, 1987, 1988a, 1988b). Such plasticity in habitat requirements holds obvious advantages from a biological control point of view as released larvae may have the potential to attack aquatic host populations in addition to utilising fresh mucus trails to locate stranded or shoreline host snails. In addition, the time before capture of the first snail meal is crucial for neonates (Beaver, 1972). Any adaptation, therefore, which facilitates prey location such as the ability to trail-follow will give species a distinct survival advantage.

Gastropod mucus is comprised of lectins, mucopolysaccharides, glycoproteins, proteins, sialic acid, urionic acid, hexosamine and a range of other molecules in an aqueous medium (Skingsley et al.,2000). Some of these compounds appear to be common to the mucus of all gastropods but others are unique to specific species. The vast majority of aquatic sciomyzids including S. s. spinipes and D. montana are generalist predators feeding on a wide range of pulmonate snails. It is likely then that the larvae of these generalist sciomyzids utilise one or more of the ubiquitous mucus components in detecting and following fresh mucus trails. It is also worth highlighting that the test animals used in this study were newly hatched larvae which had no previous contact with a snail. Hence, the ability of S. s. spinipes and D. montana neonates to follow fresh mucus trails is likely to be an innate or genetic response.

The impact of trail age on larval trail-following behaviour was also assessed. When trails were aged for 45 minutes, significantly more neonates of both species exhibited a positive response (S. s. spinipesχ2 = 9.60; D. montanaχ2 = 6.67: 1df, P < 0.01) than a non-response (Table I). However, none of the test larvae followed the mucus to its end (strong response). The majority of individuals (S. s. spinipes: 93.33%; D. montana: 86.67%) followed the trail for a short distance and then deviated from it. It appears, therefore, that the active constituent(s) is insoluble or only weakly soluble in water and that the stimulatory component in the mucus becomes inactive with time. However, an inability to effectively follow aged mucus trails may be advantageous in terms of efficient prey location. In any given snail habitat (e.g. shoreline), the substrate is likely to consist of a mosaic of mucus trails of various ages. If neonate larvae were able to follow aged mucus trails, the probability of locating prey is likely to be reduced due to incomplete (as a result of wave action for example) or overlapping trails. An acute sensitivity, however, to fresh mucus is more likely to result in successful snail location as the prey is still likely to be in the immediate environment and less energy would be expended by the larva in locating it.


Members of the Family Sciomyzidae have developed a number of different strategies to locate their prey. The females of the most highly parasitoid species such as Sciomyza aristalis (Coquillett) oviposit directly onto the larval food source (Foote 1959). Species such as Tetanocera elata (Fabricius) oviposit in the vicinity of their prey and neonate larvae remain motionless until touched by a passing slug (Knutson et al., 1965). Similar ambush behaviour is also utilised by some aquatic species (Appleton et al., 1993). It appears, therefore, as if such behaviours may have evolved to counter the patchy distribution of their gastropod prey and in instances where the larvae have to actively locate food, we postulate that some species use chemical cues to ensure a greater location rate by leading the predator to a specific resource patch. Since S. s. spinipes and D. montana are from different biogeographical zones (Palaearctic and Nearctic respectively), the ability to follow gastropod mucus trails may be a relatively common trait in larvae throughout the Family Sciomyzidae. Future research should include investigations of trail-following behaviour in species from other parts of the world, in addition to attempts at isolating the stimulatory mucus constituent (s).


Sincerest thanks to R. Orth for all his help throughout the course of this work. Many thanks also to J. Coupland, B. Foote, C. D. Williams, L. Knutson, R. Rozkošný and J.-C. Vala for their review comments and to C. Hanlon and S. Allen for technical support. We are grateful to S. McElfresh and M. Eatough-Jones for advice on statistical analysis. This research was funded in part by the European Union (MOIF-CT-2005-21592).

Copyright information

© Springer Science+Business Media, LLC 2007