All experiments were performed with the standard laboratory C. elegans strain N2 and the natural C. elegans isolate MY2079. N2 was originally obtained from the Caenorhabditis Genetics Center (CGC) and MY2079 was isolated from compost in the Botanical Garden in Kiel, Germany (Petersen et al., 2015b). All worms were maintained on nematode growth medium (NGM) plates with Escherichia coli strain OP50 as food following standard procedures (Stiernagle, 2006). Caenorhabditis elegans strains were bleached and kept for approximately 48 h on OP50 at 20°C to obtain L4 larvae or for 2 to 3 weeks at 25°C to obtain dauer larvae. All worms were washed from the plates freshly for every experiment.

Isopods, myriapods, and slugs for the experiments were caught from nature in Kiel, Germany (Supplementary Figures 1A–H). The isopod species Porcellio scaber, Oniscus asellus, and Armadillidium sp. originated from a garden plot and were mostly found below old bricks. Myriapods and slugs originated from the same compost as the natural C. elegans strain MY2079. Myriapods were sampled below tree trunks or tree bark stored close to compost. Isopods and myriapods were either directly used in experiments or kept for up to 4 days in plastic boxes with wet tissue paper and thick tree branches and leaves from their natural habitat. The wet tissue paper was replaced daily. Arion sp. slugs were kept individually for one night in boxes containing plant material from their natural habitat to reduce adhering debris.

The Drosophila melanogaster laboratory strain w[1118] was grown on standard yeast/cornmeal/agar medium at 25°C until day 4–6 of adulthood as previously described (Rahn et al., 2013; Li et al., 2015). Flies were starved in a sterile, empty fly container for 16 h at 25°C prior to the experiment with a piece of wet paper to prevent desiccation.

Washes and extracts were prepared freshly for each replicate. One isopod, one Lithobius sp. or five D. melanogaster were transferred to a sterile 1.5-ml microcentrifuge tube. Subsequently, 180 μl sterile deionized water, DMSO or ethanol were added to isopods, 250 μl to Lithobius sp. and 120 μl to D. melanogaster. Washes were obtained from invertebrates washed in the solvent with the tube rotated on an orbital rotator for 30 min. Extracts were produced from washed invertebrates ground in the solvent using a sterile pestle.

Slug feces was collected from slugs kept for 1 day in a box and placed in a 1.5-ml microcentrifuge tube. After stirring, 1 mg feces was transferred to a new tube with 180 μl deionized water, DMSO, or ethanol, homogenized with a sterile pestle, briefly centrifuged and the liquid phase transferred to a new tube. All washes and extracts were used undiluted.

We tested the chemotactic response of C. elegans to wash supernatants (washes) and invertebrates ground in deionized water, DMSO, or ethanol (extracts). All experiments were performed on 6-cm NGM plates if not stated otherwise. A letter code was used to prevent any observer bias. Petri dishes were divided into quadrants as described previously to obtain an alternating pattern of test and control to avoid any bias that might result from the initial placement of the worms (Margie et al., 2013). A circle with a diameter of 1 cm marked the center of the plate (Supplementary Figure 1I). Sodium azide (0.5 M) diluted 1:3 in all test and control liquids was used as anesthetic. Immediately after preparation of the washes and extracts, 3 μl were pipetted to opposing quadrants of the plate and 3 μl of the corresponding solvent without extract or wash to the two remaining quadrants in-between. The general response of worms to chemical stimuli was tested using 3 μl of isoamyl alcohol diluted to 10⁻³ in ethanol (attractant) or undiluted 1-octanol (repellent) as the test compound. Plates were left for 2 h to allow all liquids to sink into the agar. Finally, 30–250 C. elegans L4 or dauer larvae were added to the center, a number that allows fast processing of the assay plates, is sufficient to limit the influence of individual worms and prevents overcrowding.

Slug mucus was stamped directly from the foot of a living slug onto one side of a 9-cm agar plate. The other side of the plate remained blank. 30–250 L4 or dauer larvae were pipetted to the center of the plate.

After 1 h, worms in the test and control quadrants or halves were scored. Worms remaining in the plate center were ignored and only replicates in which at least 10 worms left the center were considered. A choice index was calculated by subtracting the number of worms in the control quadrants (or halves) from the number of worms in test quadrants (or halves) and dividing the result by the total number of worms on the plate. A positive choice index indicates the choice of the test compound (attraction), a negative choice index indicates the choice of the control (repulsion) and a choice index around zero indicates equal choice of both.

The assays were performed on unseeded 6-cm NGM plates. The center was marked with a circle of 1 cm in diameter and plates were divided into test and control halves each halve with a drilled hole in the lid. One isopod, one Lithobius sp., or five D. melanogaster were placed in a sterile 1-ml pipette tip with a 3-mm opening. The tip was placed in the lid of one halve and an identical tip without invertebrate in the other halve. All tips were sealed at the top with a heat-sterilized 8-mm cigarette filter and a piece of heat-sterilized mesh in the tip prevented escape of Lithobius sp. and D. melanogaster. The gas exchange was not actively promoted but tested using 20 μl of isoamyl alcohol diluted to 10⁻³ in ethanol (attractant) or undiluted 1-octanol (repellent) on filter paper in a pipette tip as the test odorant. Additionally, the controls were used to test the general response of worms to air-borne stimuli.

30–250 C. elegans L4 or dauer larvae were added to the plate center. After 1 h at room temperature, the tips were removed and worms on each halve were scored. A choice index was calculated as described before for the chemotaxis assays.

The experiments were performed as time-independent, biological replicates for each combination of nematode strain, nematode developmental stage, and invertebrate. Mean choice indices were reported. One-sample Wilcoxon-Signed rank tests (Wilcoxon, 1945) with false discovery rate correction (Benjamini and Hochberg, 1995) for multiple testing were performed to assess differences between a choice index of zero and the corresponding choice index of the nematode strain, developmental stage, and solvent. All statistics were performed using R-Studio (version 1.4.1717) and are presented in Supplementary Tables 1–3 and Supplementary File 1. The raw data can be found in Supplementary Tables 4–7. Composite figures were created in Inkscape (version 1.1) and Affinity Photo (version 1.10.5).

Caenorhabditis elegans’ search for hosts used for migration may potentially be facilitated by chemical signals delivered by the host to the environment. The ability to emit chemical signals is known from various invertebrates, on which C. elegans has been repeatedly found. Snails are known to release substances with their mucus that enable other snails to follow their tracks (Ng et al., 2013). Isopods such as Armadillidium vulgare use chemical signals to attract mating partners (Beauché and Richard, 2013). These chemical signals used for intra-specific communication could guide C. elegans to suitable migration partners. To test whether invertebrate secretions are attractive to C. elegans, we used wash solutions of three isopod species, as well as of myriapods, and fruit flies as potential attractants in chemotaxis assays. We found that L4 C. elegans N2 were attracted to Lithobius sp. wash in DMSO (p = 0.026), but not to wash in deionized water and ethanol. In contrast, N2 dauer larvae and neither stage of MY2079 showed attractive or repulsive behavior toward Lithobius sp. wash (Figure 1, Supplementary Tables 1, 4). Caenorhabditis elegans MY2079 dauer larvae were repelled by O. asellus wash in deionized water (p = 0.042), while L4 MY2079 and both N2 stages showed no chemotactic response to O. asellus wash. Washes of P. scaber, Armadillidium sp., and D. melanogaster did neither attract nor repel any C. elegans. Both C. elegans strains were attracted to the control attractant isoamyl alcohol and repelled by the control repellent 1-octanol as L4 larvae (Supplementary Figure 2, Supplementary Tables 1, 4). However, there was a high variability in the response of dauer larvae to both the attraction and repellent controls.

Washing dissolves only the compounds that adhere to the outside of the invertebrates. The concentration of these washed compounds might be too small to trigger a chemotactic response in C. elegans due to dilution with the solvents. Other compounds could be located inside the invertebrates, for example, in ingested food or gut content. To extract potentially attractive compounds from the inside of invertebrates, whole O. asellus, P. scaber, Armadillidium sp., Lithobius sp., and D. melanogaster were first washed in deionized water, ethanol or DMSO and then ground with a sterile pestle (Figure 2A). Caenorhabditis elegans N2 and MY2079 were not attracted to any of the ground invertebrates in either as L4 or dauer larvae (Figures 2B–F, Supplementary Tables 2, 5). In contrast, some invertebrate extracts had a repellent effect. L4 C. elegans N2 were repelled from P. scaber in deionized water (p = 0.044) and DMSO (p = 0.004), Armadillidium sp. in deionized water (p = 0.017) and ethanol (p = 0.005), and Lithobius sp. in deionized water (p = 0.005). N2 dauer larvae were neither attracted nor repelled by any invertebrate. L4 C. elegans MY2079 were repelled by O. asellus in deionized water (p = 0.016), P. scaber in deionized water (p = 0.018) and DMSO (p = 0.033), Armadillidium sp. in deionized water (p = 0.022), ethanol (p = 0.047), and DMSO (p = 0.024), and Lithobius sp. in ethanol (p = 0.034). MY2079 dauer larvae were repelled by O. asellus in ethanol (p = 0.014). L4 larvae of C. elegans N2 and MY2079 were attracted to the control attractant isoamyl alcohol and repelled by the control repellent 1-octanol (Supplementary Figure 3, Supplementary Tables 2, 5). However, dauer larvae of both C. elegans strains reacted variable to the attraction and repellent controls.

Slugs and snails have been repeatedly identified as vectors of C. elegans (Petersen et al., 2015a; Schulenburg and Félix, 2017). Since mucus plays a role in communication between snails and their conspecifics (Ng et al., 2013; Vong et al., 2019) and is therefore also a potential source of attraction for C. elegans, the attraction of mucus from the foot of Arion sp. was tested as potential attractant in the chemotaxis assay. We found that Arion sp. mucus did neither attract nor repel C. elegans (Figure 3A, Supplementary Tables 2, 6). Caenorhabditis elegans has often been isolated from the slug gut and both C. elegans and potential food bacteria can survive passage through the slug gut and be found in slug feces (Petersen et al., 2015a; Schulenburg and Félix, 2017; Pees et al., 2021). Therefore, we tested the chemotactic attraction towards slug feces next. Neither L4 nor dauer larvae of C. elegans N2 and MY2079 were attracted to Arion sp. feces (Figure 3B, Supplementary Tables 2, 6). L4 C. elegans N2 were repelled from feces in ethanol (p = 0.034). As seen before, both C. elegans strains were attracted to the control attractant isoamyl alcohol and repelled by the control repellent 1-octanol as L4 larvae (Supplementary Figure 3, Supplementary Tables 2, 6). However, the response of dauer larvae to the attraction and repellent controls was highly variable.

Entomopathogenic nematodes are attracted to volatile components of insect feces and odors emitted by live insects (Schmidt and All, 1979; Dillman et al., 2012). Carbon dioxide (CO₂) has been identified as an essential host cue for entomopathogenic nematodes (Dillman et al., 2012). Therefore, gaseous components of odorants secreted by invertebrate hosts could play a role as attractants for C. elegans. We tested this hypothesis in simple gas chemotaxis assays using live isopods, myriapods, and fruit flies. L4 C. elegans N2 were attracted to D. melanogaster (p = 0.006), but not to any other invertebrate odorant (Figure 4, Supplementary Tables 3, 7). N2 dauer larvae were not attracted to or repulsed by any invertebrate odorant. L4 and dauer larvae of C. elegans MY2079 were not attracted to or repulsed by any invertebrate odorant. Both C. elegans strains were attracted to odor of the control attractant isoamyl alcohol in as L4 larvae (Supplementary Figure 4, Supplementary Tables 3, 7). L4 larvae of N2 were repelled by the odor of the control repellent 1-octanol while C. elegans MY2079 L4 larvae showed variability in response to 1-octanol odor. Dauer larvae of both C. elegans strains showed a high variability in the response to both the attraction and repellent controls.