Do Fly Flies Detect Magnetic Fields? Evidence from a Laboratory Based Electromagnetism Experiment on T-mazes
Some of our experiments used magnetic fields that were less strong than those of Fedele et al.7 We replicated Fedele et al.7 in their original format, but did not find a magnetic response under 500 T. The radical pair mechanism12 provides no theoretical reason to expect a large difference in the responses to such similar field strengths.
Gegear et al. 2008 (ref. In an illuminated apparatus, flies experience a magnetic field generated by an electric coil system and exhibit their magnetosensitivity. The two-coil system is ideal for behavioural studies of magnetosensitivity, because it produces a magnetic field on one side of the T-maze, while producing no field on the opposite side. Importantly, the studies were carried out in the same laboratory where olfactory conditioning controls were routinely carried out in which flies are trained to associate odours with sugar reward. The wild-type flies responded to the magnetic field with naive and trained responses, and the responses were light dependent. The Cry5 photoreceptor was responsible for the light-sensitive magnetosensitivity. In a second study, Gegear et al. 2010 (ref. 3) showed that when a cry transgene is properly expressed in Cry-deficient flies, a full magnetic response with appropriate light activation is restored. The data is from published resources.
The studies published on the magnetic fields of Drosophila were dismissed because they were unable to detectmagnetic fields using a conditioning2 and negativegeot axis test. Critically, flymagnetic geotactic responses were replicated by other people. However, they aren’t discussed. Furthermore, Bae et al. successfully demonstrated a magnetic field conditioning response12, underlining how experienced Drosophila groups can successfully negotiate magnetic paradigms. I have reanalysed the data from all three geotactic experiments from Bassetto et al.1 and, despite serious flaws in methodology, their results reveal that Drosophila detect magnetic fields.
Bassetto et al.1 conclude by claiming that night-migratory songbirds (which are technically challenging for any kind of molecular genetic analyses) remain the organisms of choice for elucidating the mechanism of light-dependent magnetosensitivity. The published work on the magnetic compass of the monarch butterfly was neglected by the authors. Two independent reports that use distinctive behavioural assays show that individual monarchs manifest robust light-dependent inclination magnetic responses to Earth-strength magnetic fields11,12. Moreover, genetic studies show that the photoreceptive Cry1 protein is essential for the monarch’s light-sensitive magnetic compass12. The butterfly is an excellent choice for showing how the light-sensitive neural networks underlying light-binding in compass navigation are linked to reverse genetics.
Kyriacou1 remarks on the accuracy of the video tracking of fly movements in our negative geotaxis experiments3. The number of frames logged is simply a result of flies being out of bounds. Frames in which flies were hidden by the stoppers at the top or the supports at the base of the tubes were not included in the analysis. We also did not log flies that had arrived at the top of the tubes and had started to descend. This does not imply that flies were not tracked while they were climbing. It is not necessary to keep a record of every frame in order to determine how fast the fly is climbing. By contrast, Fedele et al.7 simply reported the proportion of flies that climbed to an arbitrarily chosen height within an arbitrarily chosen time period with no further data or photographic documentation.
The Student’s t-test (and similarly analysis of variance) proposed by Krashes and Waddell9 and used by Gegear et al.4 to analyse group T-maze data is strongly affected by pseudo-replication and is fundamentally wrong for analysing preference indices3,8. The results claimed in the ref are exaggerated. 45% naive compared to 45% trained, for small proportion contrasts. We therefore chose a correct statistical framework for proportions and avoided pseudo-replication by taking each batch of flies as the independent statistical replication unit (biological replicate). This correction, albeit conservative, nonetheless yielded significant results in our positive control experiment (odour-conditioned flies). Reppert2 includes additional statistical tests in his Comment (ordinal logistic fit model and Wilcoxon rank sum test) and presents results derived from synthetic data. The t-test has a problem that these tests have the same problem with. Statistical tests are based on frameworks of assumptions that are appropriate for specific problems and data structures and cannot be applied out of context. The original data4 and new synthetic data2 have not been made available, so we can’t comment further.
Bassetto et al.1 next reassessed the statistical analysis in Gegear et al. 2008 (ref. 2). Their reanalysis supports the contention that most of the original results were not statistically significant and were instead false positives.
An ordinal logistic fit model is equivalent to the type of generalized linear model used by Gegear et al. The results of an experiment are dependent on the number of flies in the model. The effect of training is minimal when included as an independent variable, with a noticeable effect when omitted. It is possible that the Bassetto et al.1 chose the former option.
Mimicking the approach of Bassetto et al.1 to generate a single synthetic dataset, we generated an additional 20 synthetic datasets. 5 of 20 datasets demonstrated a significant effect of training, while 15 did not. All 20 synthetic replicates demonstrated significant differences between the groups when using three other approaches. Thus, the analysis of Gegear et al.’s data seemed to have been done with a statistical approach with poor sensitivity for detecting differences. They think that the results in Gegear et al are not a false positive. False positives would occur in a range of treatments and not necessarily in a way that consistently provides evidence for magnetosensitivity, if previous studies were to be believed.
