Skip to main content
PMC home page
eLife logoLink to eLife
. 2018 Apr 13;7:e30187. doi: 10.7554/eLife.30187

Comment on "Magnetosensitive neurons mediate geomagnetic orientation in Caenorhabditis elegans"

Lukas Landler 1, Simon Nimpf 1, Tobias Hochstoeger 1, Gregory C Nordmann 1, Artemis Papadaki-Anastasopoulou 1, David A Keays 1,✉
Editor: Russ Fernald2
PMCID: PMC5898909  PMID: 29651983

Abstract

A diverse array of species on the planet employ the Earth's magnetic field as a navigational aid. As the majority of these animals are migratory, their utility to interrogate the molecular and cellular basis of the magnetic sense is limited. Vidal-Gadea and colleagues recently argued that the worm Caenorhabditis elegans possesses a magnetic sense that guides their vertical movement in soil. In making this claim, they relied on three different behavioral assays that involved magnetic stimuli. Here, we set out to replicate their results employing blinded protocols and double wrapped coils that control for heat generation. We find no evidence supporting the existence of a magnetic sense in C. elegans. We further show that the Vidal-Gadea hypothesis is problematic as the adoption of a correction angle and a fixed trajectory relative to the Earth's magnetic inclination does not necessarily result in vertical movement.

Research organism: C. elegans

Introduction

The ability to sense the Earth’s magnetic field is a widespread sensory faculty in the animal kingdom (Wiltschko and Wiltschko, 2012). Magnetic sensation has been shown in migratory birds (Zapka et al., 2009), mole rats (Nemec et al., 2001), pigeons (Keeton, 1971; Lefeldt et al., 2014; Mora et al., 2004), and turtles (Lohmann et al., 2004). While behavioral evidence supporting the existence of a magnetic sense is strong, the underlying sensory mechanisms and neuronal circuitry that transduce and integrate magnetic information are largely unknown. A major impediment to progress in the field is the lack of genetic and molecular tools in magnetosensitive species. One such model system could be the nematode Caenorhabditis elegans, which has proved to be a powerful tool to explore a wide variety of senses. It has been claimed by Vidal-Gadea et al. (2015) that C. elegans possess a magnetic sense which can easily be exploited for mechanistic investigation (see also Bainbridge et al., 2016). They argue that C. elegans possess a magnetic sense that is employed for vertical orientation, worms adopting a correction angle relative to the inclination of the Earth's magnetic field. This conclusion was based on results from three assays which they developed: (1) a ‘vertical burrowing assay’; (2) a ‘horizontal plate assay’; and (3) a ‘magnetotaxis assay’. Here, we set out to replicate the aforementioned behavioral assays, adopting several critical controls that were absent in the original study.

Results

Benzaldehyde control experiment

We established a positive control for our experiments employing the odorant benzaldehyde. It has been shown that if worms are placed in the center of a petri-dish and given the choice between 1% benzaldehyde and 100% ethanol they are attracted to the benzaldehyde. Conversely, if worms are pre-exposed to 100% benzaldehyde their preference is disrupted (Nuttley et al., 2001). Employing blinded protocols, we found that worms preferred 1% benzaldehyde (n = 11, p<0.005, Wilcoxon signed rank test), which was lost when pre-exposed to 100% benzaldehyde (Figure 1A–B). These results show that we are able to replicate published C. elegans chemotaxis experiments in our laboratory.

Figure 1. Benzaldehyde control experiment.

Figure 1.

Open in a new tab

(A) Experimental set up for the benzaldehyde-positive control experiments. Worms were placed at the release point and given a choice between 1% benzaldehyde in ethanol, or 100% ethanol. (B) Naive worms preferentially orientated toward the benzaldehyde (n = 11, p=0.002), and away from it if pre-exposed to benzaldehyde (n = 12, p=0.036). Each data point represents the result of one independent test plate.

Infrastructure and double wrapped coils

To perform the magnetic experiments described by Vidal-Gadea and colleagues we built the necessary infrastructure to insure that our experiments were performed in a clean magnetic environment. This consists of six double-wrapped Helmholtz coils, within a mu-metal shielded room that is surrounded by a Faraday cage (Figure 2A–C). Radio frequency contamination within this room is very low, with intensities below 0.1 nT between 0.1 to 10 MHz (see Figure 2A–B). This infrastructure is critical for applying magnetic stimuli in a controlled fashion (Engels et al., 2014).

Figure 2. Infrastructure for magnetic experiments.

Figure 2.

Open in a new tab

(A) All experiments were performed within a mu-metal shielded room surrounded by a 5 mm aluminum Faraday cage. DC power sources and the computer driving the Helmholtz coils were located outside this shielded room, and cables into the room were filtered for radio frequencies. (B) Graph showing the radio-frequencies present in the shielded room between 0.1 to 10 MHz are below 0.1 nT, indicative of very low levels of radio frequency contamination. (C) Experimental setup for exposure of worms to magnetic fields. Three pairs of double-wrapped Helmholtz coils surround a plastic stage in the center. Worms were placed on this stage for the vertical burrowing, horizontal plate, and magnetotaxis assays. In the burrowing assay, we surrounded the tubes by an additional small Faraday cage.

Vertical burrowing assay

In the first magnetic assay described by Vidal-Gadea, starved animals were injected into agar-filled plastic pipettes (Figure 3A). Worms were allowed to migrate overnight, and the number on each end of the tube were counted. In the absence of an external field the authors reported that animals preferentially migrated downwards, however, when exposed to an inverted Earth strength magnetic field worms migrated upwards. This preference was reversed in the case of fed animals. We repeated these experiments, but observed no effect of inverting the magnetic field on the burrowing index when the worms were starved (Mann-Whitney U-test, n1 = 38, n2 = 40, U = 681, n. s.) or fed (Mann-Whitney U-test, inclination down: n1 = 20, n2 = 35, U = 300, n. s.) (Figure 3B). The 95% confidence intervals for our experiments did not encompass the results Vidal-Gadea and colleagues obtained for the respective groups (see Supplementary file 1). Moreover, in the absence of a magnetic stimulus we found that the distribution of starved and fed worms was similar to data with an applied magentic field (Figure 3—figure supplement 1).

Figure 3. Magnetic assays and results.

(A) Diagram showing the tubes employed for the vertical burrowing assay. Worms were injected in the center hole, and NaN3 in the end-holes to immobilize them. Fed or starved worms were allowed to burrow overnight with the inclination of the magnetic field either up (59.16°) or down (−59.16°). At the conclusion of the test, the worms on either side (3 cm from the end hole) were counted and a preference index calculated. (B) Results for the vertical burrowing assay. We observed no significant difference in the burrowing index when the inclination of the magnetic field was inverted, whether the worms were fed or starved. (C) Set up for the horizontal plate assay. Worms were released in the center of the plate and allowed to move freely for 1 hr before the position and the direction of each worm relative to the center was recorded. Animals were tested in one of four magnetic directions (magnetic north pointing toward either topographic north, east, south, and west), with a field strength of 32.5 µT and 65 µT. Control experiments employed antiparallel currents resulting in a zero magnetic field. We calculated one mean orientation vector for each test plate by calculating the vector sum of all worms from this plate. (D) Results for the horizontal plate assay. We observed no directional preference when worms were exposed to either 32.5 µT or 65 µT magnetic stimuli. Each dot represents the mean worm direction for one plate, while the black arrow showing the direction and length (r) of the mean vector (radius of the circle is 1). Mag N indicates the normalized magnetic north and Topo N the topographic north. (E) Set up for the magneto-taxis assay. Worms were released in the center of a testing plate and could choose between two 3.5 cm diameter circles (goal areas) with a strong magnet (0.29 T) or a brass control underneath. Worms in each of the goal areas were counted and a preference index calculated. (F) We observed no preference for the area above the magnet, unless worms were fed bacteria contaminated with magnetite particles (p = 0.042, n = 49 plates). Error bars show standard error of the means.

Figure 3.

Figure 3—figure supplement 1. Results of the burrowing assay performed on fed and starved worms in the absence of a magnetic field.

Figure 3—figure supplement 1.

The distribution of starved and fed worms is similar.
Open in a new tab

Horizontal plate assay

In their second behavioral assay, Vidal-Gadea placed ≈50 worms in the center of an agar plate (Figure 3C). This plate was placed within a single wrapped Merritt coil system which permitted the generation of either null or horizontal magnetic fields of Earth strength intensity (either 32.5 μT or 65 μT). They reported that in the absence of magnetic stimuli worms displayed no directional preference, whereas in the presence of a horizontal field fed worms distributed in a biased direction 120° from north. We replicated these experiments, treating each plate as an experimental unit. Blind analysis of worm orientation revealed no effect on orientation behavior when applying a 32.5 µT stimulus (Rayleigh-test, r = 0.20, n = 24, n.s.) or a 65 µT stimulus (Rayleigh-test, r = 0.25, n = 24, n. s., Figure 3D). Nor did we observe any directional preference in our control experiments (32.5 µT: Rayleigh-test, r = 0.10, n = 24, n. s.; 65 µT: Rayleigh-test, r = 0.11, n = 24).

Magnetotaxis assay

In their third behavior assay, worms were placed in the center of a horizontal agar plate between two different goal areas (Figure 3E). An extremely strong neodynium magnet generating a field up to 0.29 T (approximately 8000 times Earth strength), was placed beneath one of the goal areas. Vidal-Gadea reported that in the absence of this magnet worms were distributed evenly between the goal areas, however, if the magnet was present worms migrated toward it. We replicated their set up placing a strong neodynium magnet under one goal area, but added an equally size non-magnetic brass control under the opposing goal area. We observed no preference for the goal area associated with the neodynium magnet (n = 49 plates, Wilcoxon signed rank test, V = 565, n.s., Figure 3F). The confidence interval did not include the results reported by Vidal-Gadea and colleagues (95% CI: −0.138 to 0.128). As false-positives in magnetoreception have been associated with contamination of biological material with exogenous iron we asked whether this might influence the behavior of worms (Edelman et al., 2015). We tested this by growing worms on agar plates spiked with magnetite particles, and repeated the magnetotaxis assay. We found a weak but significant preference for the goal area under which the magnet resided (Wilcoxon signed rank test, n = 49 plates, V = 670.5, p = 0.042, Figure 3F).

Discussion

Why are our results different from those of Vidal-Gadea? We have gone to great lengths to employ the same protocols. We have used worms from the same source, we have employed the same neodymium magnets, we have used the same assay plates, and the same synchronization and starvation protocols. There were, however, a number of important differences. First, we have used double wrapped coils for our experiments (Kirschvink, 1992). Our double wrapped coils (unlike single wrapped coils) allow the application of a magnetic stimulus without generating a change in temperature compared to the control condition. Heat is an issue when dealing with C. elegans as it is known that they can reliably detect temperature changes that are <0.1 °C (Ramot et al., 2008). Second, we used strict blinding procedures in all our assays, assuring an unbiased assessment of the worm responses. While Vidal-Gadea report blinding when comparing different genotypes, they do not report blinding to the magnetic condition. Third, we have applied the appropriate statistical methodology when analysing our data from the horizontal plate assay. Vidal-Gadea placed ≈50 worms on a plate treating each worm as a biological replicate. However, as worms tested on the same plate can interact with each other, they are not true independent biological replicates. The approach adopted by Vidale-Gadea is known as pseudoreplication, as it confuses the number of data points with the number of independent samples, increasing the probability of rejecting the null hypothesis whilst it is actually true (Lazic, 2010).

Moreover, there are a number of conceptual issues that undermine the assertion that C. elegans are magnetosensitive. First, the magnetotaxis assay relies on a permanent magnet that generates a field that is up to 8000 times Earth strength (0.29 T). At no time in its natural environment would C. elegans encounter such a strong field. An alternative explanation for this ‘magnetotactic behavior’ could be that exogenous iron particles attached to, or ingested by the worm, might, in the presence of an extremely large magnetic field influence the direction of locomotion by applying a force to surface mechanoreceptors.

More troubling is the underlying hypothesis that nematodes adopt a correction angle (α) relative to the inclination of the field to guide their vertical movement. Imagine a nematode is located in Cairo where the inclination of the Earth's magnetic vector is 44° 33'. To migrate vertically (i.e. 90°) it should adopt a correction angle of approximately 45° to the magnetic vector and maintain that trajectory (Figure 4A). Assuming that nematodes cannot distinguish up from down, the adoption of a fixed 45° angle from the inclination of the field is just as likely to result in horizontal movement (180°) as vertical translation (90°). This problem is exacerbated as the correction angle increases (e.g. 60°) (Figure 4B). In the best case scenario, worms could undertake random walks around a set angle (45°), that would result in a meandering descending trajectory, but with a large increase in path length. The concept proposed by Vidal-Gadea is only an efficient strategy if the worms are using the ‘correction angle’ in relation to an independent reference (i.e. gravity). However, if worms are able to distinguish up from down based on gravity, why would they rely on a magnetic field vector?

Figure 4. Conceptual issues with the Vidal-Gadea hypothesis.

Figure 4.

Open in a new tab

(A) The hypothesis advanced by Vidal-Gadea and colleagues argues that nematodes exploit the inclination of the Earth's magnetic field to guide vertical movement. They propose that nematodes adopt a correction angle (α, e.g. 45°) relative to the inclination of the field, which varies depending on the latitude. However, if the worms adopt such an angle and take a fixed trajectory this is as likely to result in a worm that travels horizontally as vertically. (B) As the latitude nears the equator the correction angle increases (e.g. 60°), and consequently a worm is just as likely to translate downwards, or at an oblique angle toward the Earth's surface. The light blue lines show the magnetic field vector.

In conclusion, we were not able to replicate the findings of Vidal-Gadea and colleagues. We have made a number of arguments why this might be the case, but it is possible that our failure to replicate this work is due to a factor we are not aware of. However, it is pertinent to note that other attempts to elicit magnetoreceptive behavior in C. elegans have also been unsuccessful (Njus et al., 2015). Collectively, we conclude that C. elegans is not a suitable model system to understand the molecular basis of magnetoreception because (a) they lack a magnetic sense, or, (b) their magnetotactic behaviour is not robust.

Methods and materials

Animals

Worms (N2 strain, received from Caenorhabditis Genetics Center) were maintained on the Escherichia coli strain OP50 as food. They were kept in incubators at constant dark conditions at 20 °C in an unmanipulated Earth-strength magnetic field (Vienna: field strength: 49 µT, inclination: 64°). For all assays, we used adult hermaphrodite worms that had not previously been starved. Worms were synchronized (bleached) before the tests to make sure animals of the same age were employed for behavioural analysis. Worms referred to as ‘fed’ were always tested within 10 mins of being removed from the culture plate. ‘Starved’ animals were kept in liquid Nematode Growth Media (NGM) for ≈30 min.

Chemotaxis experiments

For our chemotaxis experiments, we used 100 mm style petri dishes filled with 3% chemotaxis agar as test plates. Employing a template we marked each of the test plates with one center release point (see Figure 3E) and two smaller ‘scoring’ circles (diameter: 3.5 cm). Sodium azide (1.5 µl of 1 M) was applied to the center of each of the scoring circles to immobilize the worms (Nuttley et al., 2001). Worms were picked from the culture plates and collected in a small drop of NGM on a parafilm strip. In order to reduce bacterial contamination we carefully removed liquid containing bacteria and replaced it with new NGM. Worms were pipetted onto the center of the assay plate and 1 µl 1% benzaldehyde solution (in ethanol) was applied to one scoring circle and 1 µl 100% ethanol was applied to the other scoring circle. The plates were covered with aluminum foil and placed in the shielded room and left undisturbed for one hour. For our pre-exposure experiments a strip of parafilm with a 2 µl drop of 100% benzaldehyde was placed on the upper inside lid of a plate. After 90 min of pre-exposure the worms were tested as described above. For all chemotaxis experiments, we tested ≈50 worms per test. A preference index (PI) was calculated by ascertaining the difference between the number of worms reaching the benzaldehyde decision circle (B) and the 100% ethanol decision circle (E) and divided it by the total number of worms scored, PI=(B-E)/(B) + (E).

Magnetic coil set-up and magnetic shielding

For earth-strength magnetic field manipulations, we used a double wrapped custom built Helmholtz coil system (Serviciencia, S. L). The coils were located in the center of a 4.4 m (long) x 2.9 m (wide) x 2.3 m (high) shielded room. The diameter of coils were as follows: 1200 mm (Z-axis), 1254 mm (Y-axis) and 1310 mm (X-axis). The room was shielded against static magnetic fields by a 1 mm thick layer of Mu-metal and against oscillating electromagnetic fields by an aluminum layer (5 mm) (Magnetic Shielding). The ‘Inclination down’ setting as used in this study comprises a magnetic field vector with a 25 µT horizontal component, −42 µT vertical component and an inclination of −59.16°. The vertical component was inverted in the ‘inclination up’ treatment. Static magnetic fields were measured using a Three-axis Fluxgate Magnetometer (Bartington Instruments, UK). Radio frequencies were measured using an EMI test receiver (Rhode and Schwarz: MNr: E01180) and an active shielded loop antenna 6507 (EMCO: MNr: E0575). The receiver was put on MAXHOLD and measurements were taken for one min.

Burrowing assay

We used 24 cm long tubes filled with 3% chemotaxis agar (see Figure 3A), each end was closed with a plastic stopper. The tubes contained three small holes (3 mm in diameter), one in the center and two 10 cm apart from the center hole on either side. During filling of the tubes great care was taken to avoid air bubbles at the ends of the tubes. Tubes with air bubbles were discarded. 1.5 µl of 1 M NaN3 was added to each end-hole of a test tube and ≈50 were injected into the center-hole (Figure 3A). The test tube was then covered with aluminum foil and placed upright in a holder. The holder was placed in the shielded room inside a smaller copper Faraday cage (Figure 2C). Tubes were left undisturbed overnight or alternatively over a day. At the conclusion of the test the tubes were removed from the room and worms on either side (3 cm from the end hole) were counted. The ‘Inclination down’ setting as used in this study comprises a magnetic field vector with a 25 µT horizontal component, −42 µT vertical component and an inclination of −59.16°. The vertical component was inverted in the ‘inclination up’ treatment. These magnetic conditions were identical to those employed by Vidal-Gadea. We calculated the burrowing index (BI) by dividing the difference between worms on either side of the plastic tube (A), (B) by the total number of scoring worms, BI=(A-B)/(A) + (B).

Horizontal plate assay

Non-starved worms (≈50) were placed, with a droplet of NGM, on the center of a 100 mm style petri dish filled with 3% chemotaxis agar. Sodium azide (0.1 M, 20 µl) was applied to the rim of the plate to immobilize the worms once they reached it. Worms were released from the NGM droplet by removing the liquid with a tissue. The plate was then immediately placed in the center of the magnetic coils, described above, and covered with aluminum foil. Animals were tested in one of four magnetic directions (magnetic north pointing toward topographic north, east, south or west), with a field strength of 32.5 µT and 65 µT (close to the strength of the horizontal component of the Earth’s magnetic field). In addition, we used two control conditions where the double wrapped coils were switched to antiparallel currents, which resulted in a zero magnetic field. We performed this control for the 32.5 µT and 65 µT field settings. Worms were allowed to move freely on the plate for 1 hr, then the position and the direction of each worm relative to the center was recorded. Magnetic field conditions were set by a person not involved in the analysis. Treatments and field conditions were revealed after all worms were counted and the angles measured.

Magneto-taxis assay

We used 100 mm style petri dishes filled with 3% chemotaxis agar as test plates, marked with one center release point and two smaller ‘scoring’ circles. Sodium azide (1.5 µl of 1 M) was applied to the center of each of the scoring circles to immobilize the worms. We randomly placed a magnet (N42 Neodymium 3.5 cm diameter magnet 5 mm thick and nickel-plated) under one goal area, and a brass coin with identical dimensions as a control under the opposing goal area. The magnet was placed with the magnetic north pole pointing up in all tests. ≈50 worms were placed in the central release point with a droplet of NGM. After the worms were released by removing the liquid the plate was covered quickly with aluminum foil and placed in the shielded room. After 1 hr, the number of worms in each goal area were counted blind. It should be noted that Vidal-Gadea performed this experiment over 30 min; however, our pilot experiments showed that a longer time resulted in a higher percentage of worms in the goal areas. For our iron contamination experiments, the OP50 (in solution) was mixed thoroughly with magnetite to create a 1% magnetite/OP50 solution. Worms were then synchronized and grown on OP50 covered plates until they reached adulthood. Experiments were performed as described above. In order to avoid cross-contamination separate picks were used for the magnetite and non-magnetite trials. To calculate the preference index (PI) the number of worms on the magnetic side (M) were subtracted by the number of worms on the control side (C) and then divided by the total number of scoring worms, PI = (M - C)/ (M + C).

Statistics

In all tests, the experimenter was blind to the particular treatment when counting the worms. In general, we counted all tests, and did not discount tests based on low numbers of scoring worms or similar criteria in order to have an unbiased result. However, in the rare cases where no worms scored, the tests were excluded from further analysis. A one-tailed Wilcoxon one-sample test was used to test if worms preferred the benzaldehyde and the magnet. For the burrowing assay, we used a two-tailed Wilcoxon one-sample test to ascertain whether worms burrowing preference differed from zero. In order to compare groups we used a Mann–Whitney U test. All linear statistical tests were performed in R (R core team, 2012). The circular data from the horizontal plate assay were analyzed using Oriana 4. Worms tested together at the same time on the same plate can interact with each other and hence constitute non-independent samples. Therefore, we calculated one mean orientation vector for each test plate, by calculating the vector sum of all worms from this plate. The directions from the plates, relative to the magnetic field and a geographically fixed direction (door to the shielded room), were then tested for a significant unimodal orientation using the Rayleigh test. Full statistics are shown in Supplementary file 1.

Acknowledgements

We wish to thank Boehringer Ingelheim for funding basic research at the Institute for Molecular Pathology. We also wish to acknowledge the IMP graphics department and the media kitchen. We are indebted to Manuel Zimmer, Annika Nichols, and Harris Kaplan for their assistance with the experiments described in this manuscript. We thank Andres Vidal-Gadea for hosting Lukas Lander for a week in his laboratory, and showing him how to perform the magnetic behavioral protocols. The C. elegans strain was provided by the CGC which is funded by NIH Office of Research Infrastructure Programs (P40 OD 010440).

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Contributor Information

David A Keays, Email: keays@imp.ac.at.

Russ Fernald, Stanford University, United States.

Funding Information

This paper was supported by the following grants:

  • Austrian Science Fund Y726 to Lukas Landler, Simon Nimpf, Tobias Hochstoeger, Gregory Nordmann, Artemis Papadaki-Anastasopoulou, David A Keays.

  • Horizon 2020 Framework Programme 336724 to Lukas Landler, Simon Nimpf, Tobias Hochstoeger, Gregory Nordmann, Artemis Papadaki-Anastasopoulou, David A Keays.

Additional information

Competing interests

No competing interests declared.

Author contributions

Conceptualization, Supervision, Funding acquisition, Writing—original draft, Project administration, Writing—review and editing.

Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review and editing.

Data curation, Methodology.

Methodology, Writing—review and editing.

Methodology.

Methodology, Writing—review and editing.

Additional files

Supplementary file 1. Detailed summary of statistics used for the chemotaxis as well as the magnetic assays.
elife-30187-supp1.xlsx (36.2KB, xlsx)
DOI: 10.7554/eLife.30187.007
Transparent reporting form
elife-30187-transrepform.docx (250.9KB, docx)
DOI: 10.7554/eLife.30187.008

References

  1. Bainbridge C, Schuler A, Vidal-Gadea AG. Method for the assessment of neuromuscular integrity and burrowing choice in vermiform animals. Journal of Neuroscience Methods. 2016;264:40–46. doi: 10.1016/j.jneumeth.2016.02.023. [DOI] [PubMed] [Google Scholar]
  2. Edelman NB, Fritz T, Nimpf S, Pichler P, Lauwers M, Hickman RW, Papadaki-Anastasopoulou A, Ushakova L, Heuser T, Resch GP, Saunders M, Shaw JA, Keays DA. No evidence for intracellular magnetite in putative vertebrate magnetoreceptors identified by magnetic screening. PNAS. 2015;112:262–267. doi: 10.1073/pnas.1407915112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Engels S, Schneider NL, Lefeldt N, Hein CM, Zapka M, Michalik A, Elbers D, Kittel A, Hore PJ, Mouritsen H. Anthropogenic electromagnetic noise disrupts magnetic compass orientation in a migratory bird. Nature. 2014;509:353–356. doi: 10.1038/nature13290. [DOI] [PubMed] [Google Scholar]
  4. Keeton WT. Magnets interfere with pigeon homing. PNAS. 1971;68:102–106. doi: 10.1073/pnas.68.1.102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Kirschvink JL. Uniform magnetic fields and double-wrapped coil systems: improved techniques for the design of bioelectromagnetic experiments. Bioelectromagnetics. 1992;13:401–411. doi: 10.1002/bem.2250130507. [DOI] [PubMed] [Google Scholar]
  6. Lazic SE. The problem of pseudoreplication in neuroscientific studies: is it affecting your analysis? BMC Neuroscience. 2010;11:5. doi: 10.1186/1471-2202-11-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Lefeldt N, Heyers D, Schneider N-L, Engels S, Elbers D, Mouritsen H. Magnetic field-driven induction of ZENK in the trigeminal system of pigeons (Columba livia) Journal of the Royal Society Interface. 2014;11:20140777. doi: 10.1098/rsif.2014.0777. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Lohmann KJ, Lohmann CM, Ehrhart LM, Bagley DA, Swing T. Animal behaviour: geomagnetic map used in sea-turtle navigation. Nature. 2004;428:909–910. doi: 10.1038/428909a. [DOI] [PubMed] [Google Scholar]
  9. Mora CV, Davison M, Wild JM, Walker MM. Magnetoreception and its trigeminal mediation in the homing pigeon. Nature. 2004;432:508–511. doi: 10.1038/nature03077. [DOI] [PubMed] [Google Scholar]
  10. Nemec P, Altmann J, Marhold S, Burda H, Oelschlager HH. Neuroanatomy of magnetoreception: the superior colliculus involved in magnetic orientation in a mammal. Science. 2001;294:366–368. doi: 10.1126/science.1063351. [DOI] [PubMed] [Google Scholar]
  11. Njus Z, Feldmann D, Brien R, Kong T, Kalwa U, Pandey S. Characterizing the Effect of Static Magnetic Fields on C. elegans Using Microfluidics. Advances in Bioscience and Biotechnology. 2015;6:583–591. doi: 10.4236/abb.2015.69061. [DOI] [Google Scholar]
  12. Nuttley WM, Harbinder S, van der Kooy D. Regulation of distinct attractive and aversive mechanisms mediating benzaldehyde chemotaxis in Caenorhabditis elegans. Learning & Memory. 2001;8:170–181. doi: 10.1101/lm.36501. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. R core team . Vienna , Austria: R Foundation for Statistical Computingatistical Computing; 2012. [Google Scholar]
  14. Ramot D, MacInnis BL, Goodman MB. Bidirectional temperature-sensing by a single thermosensory neuron in C. elegans. Nature Neuroscience. 2008;11:908–915. doi: 10.1038/nn.2157. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Vidal-Gadea A, Ward K, Beron C, Ghorashian N, Gokce S, Russell J, Truong N, Parikh A, Gadea O, Ben-Yakar A, Pierce-Shimomura J. Magnetosensitive neurons mediate geomagnetic orientation in Caenorhabditis elegans. eLife. 2015;4:e07493. doi: 10.7554/eLife.07493. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Wiltschko R, Wiltschko W. Magnetoreception. In: López-Larrea C, editor. In Sensing in Nature. New York, NY: Springer US; 2012. pp. 126–141. [DOI] [Google Scholar]
  17. Zapka M, Heyers D, Hein CM, Engels S, Schneider NL, Hans J, Weiler S, Dreyer D, Kishkinev D, Wild JM, Mouritsen H. Visual but not trigeminal mediation of magnetic compass information in a migratory bird. Nature. 2009;461:1274–1277. doi: 10.1038/nature08528. [DOI] [PubMed] [Google Scholar]

Decision letter

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "Comment on "Magnetosensitive neurons mediate geomagnetic orientation in Caenorhabditis elegans" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Eve Marder as the Senior Editor. The following individual involved in review of your submission has agreed to reveal his identity: Markus Meister (Reviewer #1).

Summary:

There is general agreement that the biological basis for a magnetic sense is one of the last great frontiers in sensory biology. The vexed history of magnetodetection in many other systems supports the view that C. elegans is an excellent system for understanding magnetodetection at the cellular and molecular levels. That is why we welcome your challenge to the research of Vidal-Gadea and colleagues, as we all agree that it is important to determine the validity of the original study.

General comments:

1) Subsection “Vertical Burrowing Assay”: "We repeated these experiments, but observed no effect of inverting the magnetic field on the burrowing index when the worms were starved…(n.s.)": The authors' results are much stronger than stated here and their presentation should be revised. (1) The statement currently focuses on "no effect of inverting the field" but actually the observation is simply "no effect at all". There is no evidence for directional burrowing, i.e. a non-zero BI, under any of the conditions. (2) For making a negative claim it is not sufficient to merely state "no significant difference". You can get "no significant difference" simply by not experimenting hard enough. Absence of significance is not significance of absence. A better method (not the only one) is to state the 4 results as mean {plus minus} 95% confidence limits. Then interpret this by showing that the null hypothesis (BI = 0) is within the confidence limits in all 4 cases, and that the numbers reported by Vidal-Gadea Figure 1B and 2G are far outside the confidence limits.

2) “Horizontal Plate Assay”: Again, statistical significance is irrelevant for making your case. Please state the strength of the directionality effect with confidence limits. Test whether that is consistent with zero effect and whether it is consistent with the Vidal-Gadea’s claims. In this case the effects in Vidal-Gadea Figure 2 are rather weak and may well be consistent with your measurements.

3) “Magnetotaxis Assay”: Again, the p-value is not helpful. State the effect size with confidence limits and compare to the Null hypothesis and to Vidal-Gadea Figure 4.

4) Introduction section: "Magnetic sensation has been shown in…". The authors may want to express this more carefully, especially given their current pursuit. Several of the papers they cite here contain weaker behavioral evidence than the Vidal-Gadea 2015 report. Wu and Dickman 2012 contains strong data, but has never been replicated, not even in the PI's laboratory, and not for lack of trying.

5) Subsection “Infrastructure and double wrapped coils”: "Radio frequency contamination within this room is very low…" What hypothesis would lead one to be interested in radio frequency magnetic fields? If you want to reassure us about possible environmental confounds, I would be more interested in odor plumes for example.

6) Subsection “Horizontal Plate Assay”: "worms distributed in a biased direction 60° either side of the imposed vector.": Actually in Vidal-Gadea Figure 2, the direction is 120 degrees from North for fed worms, and 60 degrees from North for starved worms.

7) Subsection “Magnetotaxis Assay”: "was placed beneath one of the goal areas": Strictly speaking Vidal-Gadea placed the magnet "above" the goal area.

8) Discussion section, paragraph three: More generally, worms should with equal probability choose directions on a cone whose surface is 45 degrees relative to the magnetic axis. On average they will therefore travel in the direction of the magnetic axis, but slower by a factor of 1/sqrt(2). So if the goal is to descend vertically, this strategy will take 41% longer on average than simply following the field.

9) Materials and methods section: Define "OP50"?

10) Figure 1: Maybe state that each dot is an independent test involving >50 worms.

11) Figure 2 legend: Is Figure 2B a power spectrum? If so, please state and label vertical axis accordingly.

12) I understand the authors' argument regarding the use of large double-wrapped coils instead of the single-wrapped coils used by Vidal-Gadea et al. However, it would be valuable for the authors to repeat the Vidal-Gadea experiments exactly according to their protocols, to see if this difference (and presumably generated heat) is responsible for the different results. (While I agree with the authors' points about double blinding and biological replicates, neither of these factors seems likely to explain the very divergent experimental outcomes.)

13) On a similar note, in the acknowledgements, it is stated that one of the authors (LL) visited the Vidal-Gadea lab. Did LL try to replicate the magnetotaxis results when he was there? Or did this visit simply involve an observation of protocols? Again, this might address specifically why the authors obtained different results from those in the Vidal-Gadea paper. At minimum, the authors should provide more detail about what the visits involved.

14) While the data presented challenge the results published by Vidal-Gadea and colleagues, we strongly recommend more modest interpretation of the data, especially omission of the last sentence of the paper. The failure of the replication cannot automatically be used to disprove original study. History of magnetoreception research teaches us that replication of some experiments may be difficult. For instance, it has taken years to replicate independently original experiments showing magnetic compass orientation in birds.

15) One potential caveat of the study is that the worms were kept and tested in different magnetic conditions. If I am right, the worms were kept in normal (i.e., rather noisy) magnetic field in a laboratory and then tested in the magnetically shielded Mu-metal chamber. It has been shown repeatedly in birds that pre-exposure to the same magnetic intensity and light conditions may be necessary for efficient magnetic orientation. For the sake of reproducibility and critical assessment of study design, magnetic and light conditions to which animals were exposed before and during the experiments must be described in detail.

16) The authors indicate levels of radiofrequency fields in the shielded experimental room between 0.5 and 5 MHz. Because it has been shown that broadband radiofrequencies (0.1 – 10 MHz) and single frequency of 7 MHz disrupt magnetic compass orientation of birds (Ritz et al., 2004), much broader frequency range has to be characterized to exclude possible interference with radical-pair mechanism. Information concerning levels of ELMF before and during experiment shall also be given.

17) While iron-contamination from the laboratory may lead to false positive results, the way in which the experiment was done seems unrealistic. The worms were grown on 1% magnetite/OP50 bacteria solution. Such a strong contamination is highly unlikely under standard lab conditions. In this context, the claim made in the Abstract (we … demonstrates that iron-contamination from laboratory settings can result in false positive results) is inappropriate and must be reformulated.

18) In the Discussion, the authors claim: "Our large double wrapped coils allow the application of a magnetic stimulus without generating a change in temperature compared to the control condition. In contrast, the small single wrapped coils employed by Vidal-Gadea generate a temperature gradient." This is a misleading statement for two reasons. First, Vidal-Gadea and colleagues used Merrit coil of 1 m3. It is not small for experiments with C. elegans that are performed on 100 mm petri dishes. Second, Vidal-Gadea et al. reported some measurable temperature gradient between center and periphery of petri dishes. However, there was no significant difference in this temperature gradient between different magnetic conditions (see their Figure 2—figure supplement 1). Thus, they actually did not face problem that your double wrapped coil solves. On the other hand, double wrapped coil in both parallel and anti-parallel mode will produce some heat and therefore potentially create temperate gradient on plates. Indicate size of Helmholz coils used in the study and temperature measurements if available.

19) The authors discuss problems associated with utilization of the inclination of the Earth magnetic vector as a cue for vertical movement. While the inclination is not an unambiguous cue, I do not agree that an independent reference (e.g. gravity) is needed. The inclination of magnetic vector is good enough for "directional random walk", which may suffice navigational needs of C. elegans.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Comment on "Magnetosensitive neurons mediate geomagnetic orientation in Caenorhabditis elegans" for further consideration at eLife. Your revised article has been favorably evaluated by Eve Marder (Senior editor), a Reviewing editor, and three reviewers.

The manuscript has been improved but there are some remaining issues that need to be addressed before acceptance, as outlined below:

General Comments:

The authors have generally done a good job responding to reviewer comments. We agree with their view that it is important to report instances in which earlier work is difficult or impossible to replicate, and the reviewers share your skepticism about aspects of the Vidal-Gadea paper, particularly the idea that worms use the earth's magnetic field to orient vertically in the soil.

Subsection “Infrastructure and double wrapped coils” and point 5 of the original review. Please add a citation of Engels, 2014 or some other explanation why you are measuring RF fields.

Subsection “Vertical Burrowing Assay”: Could the authors show control measurements without a magnetic field, and test whether the variance in worm trajectories is greater with than without field. This would address the concern that the assay may have averaged over fed and starved worms with positive and negative magnetotaxis.

Subsection “Magnetotaxis Assay”: "under one goal area", point 7 of the original review. To the reader this appears as a departure from Vidal-Gadea's published protocol "above one goal area". Please add the explanation you give in the reply to reviewers.

Discussion section, paragraph two: As pointed out in Vidal-Gadea's rebuttal, this is true only if the magnet were to lie in the plane of the worms. In reality the magnet is somewhat below that plane and that asymmetry produces a horizontal component to the field. Also in Figure 4A the field lines are unrealistic, far exaggerating the vertical component. This should be corrected or, perhaps better, the panel and its associated argument could be omitted.

Discussion section, paragraph three: "spiraling descending trajectory": I don't think a random walk would lead to any kind of spiral. But regardless of the overall shape of the path, if the worm always travels at 45 degrees relative to the intended direction, its rate of progress in that direction will be only 71% of optimum.

The first paragraph of the Discussion seems to attribute the lack of replication to one of three specific defects in Vidal-Gadea et al's experimental design. Without actually testing single wrapped coils in their hands, the authors cannot conclude that this is the causative factor. Likewise, I think the behavioral assay (counting paralyzed worms in a circle) should be relatively resistant to observer bias, and Vidal-Gadea's results look different from the author's irrespective of the statistical test used. It is quite possible, even likely, that a different factor explains the different results, and this should be noted in the Discussion.

eLife. 2018 Apr 13;7:e30187. doi: 10.7554/eLife.30187.012

Author response

General comments:

1) Subsection “Vertical Burrowing Assay”: "We repeated these experiments, but observed no effect of inverting the magnetic field on the burrowing index when the worms were starved…(n.s.)": The authors' results are much stronger than stated here and their presentation should be revised. (1) The statement currently focuses on "no effect of inverting the field" but actually the observation is simply "no effect at all". There is no evidence for directional burrowing, i.e. a non-zero BI, under any of the conditions. (2) For making a negative claim it is not sufficient to merely state "no significant difference". You can get "no significant difference" simply by not experimenting hard enough. Absence of significance is not significance of absence. A better method (not the only one) is to state the 4 results as mean {plus minus} 95% confidence limits. Then interpret this by showing that the null hypothesis (BI = 0) is within the confidence limits in all 4 cases, and that the numbers reported by Vidal-Gadea Figure 1B and 2G are far outside the confidence limits.

We have added the 95% confidence intervals to our statistical analysis for the burrowing index (See Supplementary file 1). In all cases the mean results obtained by Vidal-Gadea and colleagues were outside our 95% confidence intervals. We now state this in subsection “Vertical Burrowing Assay”. The null hypothesis (BI=0) fell within the confidence interval in all cases, with the exception of worms that are starved and presented with a downward inclination field. In this case the lower confidence interval is 0.006, reflecting the fact that starved worms in general tend to burrow upwards. A Mann-Whitney U-test test confirmed the absence of a magnetic effect.

2) “Horizontal Plate Assay”: Again, statistical significance is irrelevant for making your case. Please state the strength of the directionality effect with confidence limits. Test whether that is consistent with zero effect and whether it is consistent with the Vidal-Gadea claims. In this case the effects in Vidal-Gadea Figure 2 are rather weak and may well be consistent with your measurements.

In the case of circular statistics the use of confidence intervals when the data distribution is wide is problematic. For instance, when applying a 32.5 µT field our confidence interval varied from 57.8° to 213.3°representing 43% of the possible variation in the dataset. Accordingly the likelihood that another dataset falls within this confidence interval is high, and not informative. We have nonetheless added confidence intervals to Supplementary file 1, but think they should be interpreted cautiously.

3) “Magnetotaxis Assay”: Again, the p-value is not helpful. State the effect size with confidence limits and compare to the Null hypothesis and to Vidal-Gadea Figure 4.

We now present the confidence intervals for the magnetotaxis assay in Supplementary file 1, and in the main text. The results reported by Vidal-Gadea for the N2 strain fall outside these confidence intervals – which we now state in the text.

4) Introduction section: "Magnetic sensation has been shown in…". The authors may want to express this more carefully, especially given their current pursuit. Several of the papers they cite here contain weaker behavioral evidence than the Vidal-Gadea 2015 report. Wu and Dickman, 2012 contains strong data, but has never been replicated, not even in the PI's laboratory, and not for lack of trying.

We appreciate that the Wu and Dickman electrophysiological studies have yet to be independently replicated, however, we think there is good evidence for a magnetic sense in pigeons. To address the reviewers’ concerns we now cite (Lefeldt et al., 2014), (Mora et al., 2004), and (Keeton, 1971).

5) Subsection “Infrastructure and double wrapped coils”: "Radio frequency contamination within this room is very low…" What hypothesis would lead one to be interested in radio frequency magnetic fields? If you want to reassure us about possible environmental confounds, I would be more interested in odor plumes for example.

Radio frequency noise at nT levels has been shown to disrupt the magnetic compass of animals, e.g. in migratory robins (Engels et al., 2014) presumably because of an underlying radical pair mechanism. Therefore RF shielding and measurements are common and important practice in magnetic orientation experiments. Odor plumes are an unlikely confounding factor in our case as the experiments were done using closed plates, or sealed tubes, in an enclosed room.

6) Subsection “Horizontal Plate Assay”: "worms distributed in a biased direction 60° either side of the imposed vector.": Actually in Vidal-Gadea Figure 2, the direction is 120 degrees from North for fed worms, and 60 degrees from North for starved worms.

We have clarified this in the text of our paper. We now state that "in the presence of a horizontal field fed worms distributed in a biased direction 120° from North".

7) Subsection “Magnetotaxis Assay”: "was placed beneath one of the goal areas": Strictly speaking Vidal-Gadea placed the magnet "above" the goal area.

In the supplementary material of their manuscript Vidal-Gadea et al., 2015 state that the magnet was placed “above one of the circles”. In initial preliminary trials we also performed the experiments this way. However, when LL visited the Vidal-Gadea lab, he was informed that in reality they perform their experiments by putting the plates on top of the magnets. We therefore performed the experiment in this way.

8) Discussion section, paragraph three: More generally, worms should with equal probability choose directions on a cone whose surface is 45 degrees relative to the magnetic axis. On average they will therefore travel in the direction of the magnetic axis, but slower by a factor of 1/sqrt(2). So if the goal is to descend vertically, this strategy will take 41% longer on average than simply following the field.

We appreciate the reviewers’ comment. Having given this matter some thought we think this depends on whether or not an individual worm adopts a fixed direction with respect to the magnetic vector. If the worm employs a "random walking" approach we think the reviewer is correct and you could expect them to descend around that vector in a spiraling fashion but at a slower rate. Alternatively, if the animal adopts a single fixed heading and maintains that vector it could either be going horizontally or vertically. To address this issue we have added the following sentence to the manuscript "In the best case scenario worms could undertake random walks around a set angle (45°), that would result in a spiraling descending trajectory, but with a large increase in path length."

9) Materials and methods section: Define "OP50"?

OP50 is the standard bacterial lawn (E. coli) used to feed C. elegans. We have clarified this in the Materials and methods.

10) Figure 1: Maybe state that each dot is an independent test involving >50 worms.

We have stated this in the figure legend.

11) Figure 2 legend: Is Figure 2B a power spectrum? If so, please state and label vertical axis accordingly.

In our view the axes are labeled correctly showing the maximum strength of the magnetic field (B, in nT) on the vertical axis and the frequency in MHz on the horizontal axis.

12) I understand the authors' argument regarding the use of large double-wrapped coils instead of the single-wrapped coils used by Vidal-Gadea et al. However, it would be valuable for the authors to repeat the Vidal-Gadea experiments exactly according to their protocols, to see if this difference (and presumably generated heat) is responsible for the different results. (While I agree with the authors' points about double blinding and biological replicates, neither of these factors seems likely to explain the very divergent experimental outcomes.)

We understand the reviewers’ argument, however, our goal was to ascertain whether or not C. elegans are a useful model system to study magnetoreception. We think there is little to be gained from investing time and effort in replicating experiments with obvious design flaws. To do so would require us to construct a new set of coils and re-do an entire set of experiments. Moreover, even if we adopted this course it is unlikely we could build an exact replica of the coils employed by Vidal-Gadea.

13) On a similar note, in the acknowledgements, it is stated that one of the authors (LL) visited the Vidal-Gadea lab. Did LL try to replicate the magnetotaxis results when he was there? Or did this visit simply involve an observation of protocols? Again, this might address specifically why the authors obtained different results from those in the Vidal-Gadea paper. At minimum, the authors should provide more detail about what the visits involved.

LL visited the Vidal-Gadea lab for one week and performed magnetotaxis and burrowing tests during this week. The primary goal of this visit was to learn and observe the behavioural protocols developed by Vidal-Gadea. An analysis of the results obtained did not reveal a significant magnetic effect, but the n number was low, and the experiments were not performed in shielded conditions and not always blinded. We have added information regarding the visit to the acknowledgements.

14) While the data presented challenge the results published by Vidal-Gadea and colleagues, we strongly recommend more modest interpretation of the data, especially omission of the last sentence of the paper. The failure of the replication cannot automatically be used to disprove original study. History of magnetoreception research teaches us that replication of some experiments may be difficult. For instance, it has taken years to replicate independently original experiments showing magnetic compass orientation in birds.

We are acutely aware of the difficulty of behavioral experiments in the field magnetosensation. We think this needs to be balanced with the need for independent replication of key papers. We do not, as stated by the reviewer claim to have "disproved" the original study. Proof is an ellusive and difficult goal to obtain, requiring (in our view) multiple independent labs that perform experiments that stand the test of time.

Rather our argument is that worms are not a good model to study magnetoreception. We have clarified this by modifying the last sentence so it now reads "Collectively, we conclude that C. elegans is not a suitable model system to understand the molecular basis of magnetoreception because (a) they lack a magnetic sense, or, (b) their magnetotactic behaviour is not robust."

15) One potential caveat of the study is that the worms were kept and tested in different magnetic conditions. If I am right, the warms were kept in normal (i.e., rather noisy) magnetic field in a laboratory and then tested in the magnetically shielded Mu-metal chamber. It has been shown repeatedly in birds that pre-exposure to the same magnetic intensity and light conditions may be necessary for efficient magnetic orientation. For the sake of reproducibility and critical assessment of study design, magnetic and light conditions to which animals were exposed before and during the experiments must be described in detail.

As requested we have added details to the methodology. Specifically, worms were kept in complete darkness in incubators which were kept constant at 20°C in the undisturbed Earth’s magnetic field in Vienna (field strength: 49µT, inclination: 64°). All experiments were performed in darkness.

16) The authors indicate levels of radiofrequency fields in the shielded experimental room between 0.5 and 5 MHz. Because it has been shown that broadband radiofrequencies (0.1 – 10 MHz) and single frequency of 7 MHz disrupt magnetic compass orientation of birds (Ritz et al., 2004), much broader frequency range has to be characterized to exclude possible interference with radical-pair mechanism. Information concerning levels of ELMF before and during experiment shall also be given.

In light of the reviewers’ request we have now included an expanded spectrum showing the radiofrequency fields in our mu metal shielded room. The frequency range now extends from 0.1 MHz to 10MHz, with intensities still below 0.1nT. We did not obtain spectra during the experiments.

17) While iron-contamination from the laboratory may lead to false positive results, the way in which the experiment was done seems unrealistic. The worms were grown on 1% magnetite/OP50 bacteria solution. Such a strong contamination is highly unlikely under standard lab conditions. In this context, the claim made in the Abstract (we.. . demonstrates that iron-contamination from laboratory settings can result in false positive results) is inappropriate and must be reformulated.

We have removed this sentence from the Abstract.

18) In the Discussion, the authors claim: "Our large double wrapped coils allow the application of a magnetic stimulus without generating a change in temperature compared to the control condition. In contrast, the small single wrapped coils employed by Vidal-Gadea generate a temperature gradient." This is a misleading statement for two reasons. First, Vidal-Gadea and colleagues used Merrit coil of 1 m3. It is not small for experiments with C. elegans that are performed on 100 mm petri dishes. Second, Vidal-Gadea et al. reported some measurable temperature gradient between center and periphery of petri dishes. However, there was no significant difference in this temperature gradient between different magnetic conditions (see their Figure 2—figure supplement 1). Thus, they actually did not face problem that your double wrapped coil solves. On the other hand, double wrapped coil in both parallel and anti-parallel mode will produce some heat and therefore potentially create temperate gradient on plates. Indicate size of Helmholz coils used in the study and temperature measurements if available.

We removed the statement regarding the size of the coils employed by Vidal-Gadea and added our coil size to the methods. It is true that our double-wrapped coils may produce some heat. Critically, however, any heat generated will be identical when comparing antiparallel and parallel conditions as the exact same amount of current is flowing in both cases. This cannot be said for single wrapped coils.

19) The authors discuss problems associated with utilization of the inclination of the Earth magnetic vector as a cue for vertical movement. While the inclination is not an unambiguous cue, I do not agree that an independent reference (e.g. gravity) is needed. The inclination of magnetic vector is good enough for "directional random walk", which may suffice navigational needs of C. elegans.

We agree with the reviewer that the inclination of the magnetic field alone would be a sufficient cue to move up or down (e.g. magnetotactic bacteria). However, the hypothesis advanced by Vidal-Gadea is that the worms adopt a correction angle relative to this vector (e.g. 45 degrees). If a worm follows a straight path, adopting a fixed correction angle, it will not necessarily result in vertical translocation – it is just as likely to end up travelling horizontally. However, we do acknowledge that the adoption of a "random walking" approach around an inclination angle would result in a worm that spirals downward (with a greater path length). We have modified the manuscript accordingly to reflect this.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

General Comments:

The authors have generally done a good job responding to reviewer comments. We agree with their view that it is important to report instances in which earlier work is difficult or impossible to replicate, and the reviewers share your skepticism about aspects of the Vidal-Gadea paper, particularly the idea that worms use the earth's magnetic field to orient vertically in the soil.

We are thankful for the positive assessment of our work, and feel that the additional changes which we have done in this round of revisions have further improved our manuscript.

Subsection “Infrastructure and double wrapped coils” and point 5 of the original review. Please add a citation of Engels, 2014 or some other explanation why you are measuring RF fields.

We added the requested citation.

Subsection “Vertical Burrowing Assay”: Could the authors show control measurements without a magnetic field, and test whether the variance in worm trajectories is greater with than without field. This would address the concern that the assay may have averaged over fed and starved worms with positive and negative magnetotaxis.

We have added these data (See Figure 3—figure supplement 1). We find that in the absence of a magnetic field the variance in trajectories when comparing starved and fed worms is similar.

Subsection “Magnetotaxis Assay”: "under one goal area", point 7 of the original review. To the reader this appears as a departure from Vidal-Gadea's published protocol "above one goal area". Please add the explanation you give in the reply to reviewers.

When LL visited the Vidal-Gadea Lab he was specifically shown to put the magnets underneath the plates, therefore we adopted the same method.

Discussion section, paragraph two: As pointed out in Vidal-Gadea's rebuttal, this is true only if the magnet were to lie in the plane of the worms. In reality the magnet is somewhat below that plane and that asymmetry produces a horizontal component to the field. Also in Figure 4A the field lines are unrealistic, far exaggerating the vertical component. This should be corrected or, perhaps better, the panel and its associated argument could be omitted.

In light of the reviewers’ comments we have omitted the panel and associated argument.

Discussion section, paragraph three: "spiraling descending trajectory": I don't think a random walk would lead to any kind of spiral. But regardless of the overall shape of the path, if the worm always travels at 45 degrees relative to the intended direction, its rate of progress in that direction will be only 71% of optimum.

We have changed the wording on this paragraph indicating the worms would undertake a "meandering descending trajectory"

The first paragraph of the Discussion seems to attribute the lack of replication to one of three specific defects in Vidal-Gadea et al's experimental design. Without actually testing single wrapped coils in their hands, the authors cannot conclude that this is the causative factor. Likewise, I think the behavioral assay (counting paralyzed worms in a circle) should be relatively resistant to observer bias, and Vidal-Gadea's results look different from the author's irrespective of the statistical test used. It is quite possible, even likely, that a different factor explains the different results, and this should be noted in the Discussion.

We now state "In conclusion, we were not able to replicate the findings of Vidal-Gadea and colleagues. We have made a number of arguments why this might be the case, but it is possible that our failure to replicate this work is due to a factor we are not aware of."

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    Supplementary file 1. Detailed summary of statistics used for the chemotaxis as well as the magnetic assays.
    elife-30187-supp1.xlsx (36.2KB, xlsx)
    DOI: 10.7554/eLife.30187.007
    Transparent reporting form
    elife-30187-transrepform.docx (250.9KB, docx)
    DOI: 10.7554/eLife.30187.008

    Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd

    RESOURCES