A geomagnetic declination compass for horizontal orientation in fruit flies
Abstract
The Earth's geomagnetic field (GMF) is known to act as a sensory cue for magnetoreceptive animals such as birds, sea turtles, and butterflies in long-distance migration, as well as in flies, cockroaches, and cattle in short-distance movement or body alignment. Despite a wealth of information, the way that GMF components are used and the functional modality of the magnetic sense are not clear. A GMF component, declination, has never been proven to be a sensory cue in a defined biological context. Here, we show that declination acts as a compass for horizontal food foraging in fruit flies. In an open-field test, adopting the food conditioning paradigm, food-trained flies significantly orientated toward the food direction under ambient GMF and under eastward-turned magnetic field in the absence of other sensory cues. Moreover, a declination change within the natural range, by alteration only of either the east–west or north–south component of the GMF, produced significant orientation of the trained flies, indicating that they can detect and use the difference in these horizontal GMF components. This study proves that declination difference can be used for horizontal foraging, and suggests that flies have been evolutionarily adapted to incorporate a declination compass into their multi-modal sensorimotor system.
Introduction
The total intensity and inclination have been known to be the crucial components of geomagnetic field (GMF) that are used for horizontal migration in magnetoreceptive creatures including birds, sea turtles, flies, and butterflies (Bazalova et al. 2016; Begall et al. 2008; Gegear et al. 2008; Guerra et al. 2014; Johnsen & Lohmann 2005; Merlin et al. 2012; Wiltschko & Wiltschko 2006). Declination, a GMF component, varies marginally by either the time of day or season at a given location and is significantly different between areas (Courtillot & Le Mouel 1988; Finlay et al. 2010; Liboff 2014). It is notably more variable than total intensity and inclination due to both terrestrial and extra-terrestrial factors. Indeed, isolines of declination, the east–west component, and the north–south component of the GMF, are substantially non-linear and complex on the surface of Earth, because of local anomalies produced by mountains, rocks, and even man-made architecture (Finlay et al. 2010; Thébault et al. 2015). Even though declination was suggested to affect the heading of some birds under particular artificial magnetic environments several decades ago (Emlen et al. 1976; Wiltschko & Wiltschko 1972), it has not been possible to properly evaluate whether declination is a genuine cue for magnetoreceptive behavior, because the declination value was either not noted or set to a constant value of 0° in most of the subsequent magnetoreception studies. Recently, dogs have been observed to be sensitive to the rate of change in declination for their body alignment during defecation and urination, implying that declination might be a sensory cue (Hart et al. 2013). Cockroaches showed significant spontaneous body turns, a magnetosensitive directional response, caused by repeated directional change in the horizontal component of the GMF (Bazalova et al. 2016). However, no previous study has demonstrated the role of declination in a functionally defined biological context. Therefore, investigation of declination as a genuine sensory cue for magnetoreceptive behavior is lacking.
The fruit fly, Drosophila melanogaster, is an agile magnetosensitive creature that can use either the GMF or a static magnetic field as a sensory cue for its horizontal or vertical movements (Bae et al. 2016; Fedele et al. 2014; Gegear et al. 2008; Painter et al. 2013). Fruit flies can travel up to ten kilometers in a few hours, with various spatial scales for foraging, mating, and avoidance of predators, and their dispersal tendencies are heterogeneous and non-random (Chow & Frye 2009; Coyne et al. 1982; Slatkin 1985). Visual and olfactory cues are believed to be efficiently integrated into the sensory-motor system; for example, to locate food in global and local searches. Yet it has been proposed that a sensory-independent search capacity would be necessary in the frequent absence of olfactory and visual cues (Chow & Frye 2009). However, there is no convincing evidence to support the idea of this sensory-independent search capacity, or that there is any alternative sensory cue to compensate in the absence of usable visual-olfactory cues. We postulated that declination – a component of consistent GMF, which is variable in a range between areas – could be a reliable sensory cue for securing food without the use of visual-olfactory cues.
Materials and methods
Fly stocks
The flies were reared as described in our previous study (Bae et al. 2016) with a standard cornmeal–yeast–agar diet at 25 ± 0.5°C, 60 ± 2 % relative humidity, in a 12 h light/12 h dark cycle under a full-spectrum fluorescent light (350–800 nm; light intensity: 4.22 × 1022 photons/cm2/s) that was turned on at 09:00 (local time). At the rearing site, the ambient geomagnetic field (GMF) had a total intensity = 44.9 μT, X (north–south) = 32.2 μT, Y (east–west) = −5.8 μT, Z (vertical to ground) = 30.8 μT. The Canton-S fly strain was provided by the Bloomington Drosophila Stock Center (Indiana University, Bloomington, IN, USA). Both male and female adult flies were used together in the experiments.
Modulation of MF
A rectangular Helmholtz coil system modified from our previous studies (Bae et al. 2013; Ryu et al. 2009), consisting of three pairs of parallel coils arranged orthogonally for three axes, was used to generate static magnetic field (MF) for experimental conditions as described in our previous study (Bae et al. 2016). Briefly, the average dimensions of the coils were 1,890 × 1,890 mm, 1,800 × 1,800 mm, and 1,980 × 1,980 mm, for the X, Y, and Z axes respectively. We intentionally set the coil for the X-axis (north–south) to be aligned with geographic north so that the Y-coil (east–west axis) had room for modulating the Y component of the MF. The direction of geographic north was determined using the local declination value (−7.53° at the time of setup) of Daegu City where the experiments were performed. A pair of coils for each axis was connected to a DC power supply (E3631A; Agilent Technologies, Santa Clara, CA, USA). GMF was measured using a 3-axis gaussmeter (MGM 3AXIS; ALPHALAB, West Salt Lake City, UT, USA), and the homogeneity of the MFs in the sample space was measured as 99 %. The ambient GMF and the modulated MFs parameters are indicated in Table 1. Across the assay area and throughout the experiments, alternating MFs and electric fields up to 1 MHz were less than 3 μT and 1.20 V/m, respectively, as measured by an electromagnetic field analyzer (3D NF Analyzer NFA 1000; Gigahertz Solutions, Fürth, Bayern, Germany). All the experiments using the coil were performed in a temperature-controlled room, and the temperature across the assay area was measured as 25 ± 0.3°C using thermometers (USB Data logger 98581; MIC Meter Industrial Company, Taichung City, Taiwan).
| MF | D (°) | X (μT) | Y (μT) | Z (μT) | H (μT) | F (μT) | I (°) |
|---|---|---|---|---|---|---|---|
| Ambient GMF | −5.71 | 31.61 | −3.16 | 32.18 | 31.77 | 45.22 | 45.37 |
| Eastward-turned MF | 84.29 | 3.16 | 31.61 | 32.18 | 31.77 | 45.22 | 45.37 |
| MF 1 | 14.03 | 23.69 | 5.92 | 37.12 | 24.42 | 44.43 | 56.66 |
| MF 2 | −14.03 | 23.69 | −5.92 | 37.12 | 24.42 | 44.43 | 56.66 |
| MF 3 | 14.03 | −23.69 | −5.92 | 37.12 | 24.42 | 44.43 | 56.66 |
- MF, magnetic field; GMF, geomagnetic field; D, declination; X, north–south component; Y, east–west component; Z, vertical component; H, horizontal component; F, total intensity; I, inclination; ambient GMF, the GMF in the laboratory where experiments were conducted; eastward-turned MF, the MF at 90° clockwise from the ambient magnetic north. MFs 1, 2, and 3: the MFs for the comparative orientation tests under paired MFs, MFs 1 & 2 and MFs 2 & 3.
Horizontal orientation assay
To assess horizontal orientation of fruit flies using a food conditioning paradigm, an open-field arena comprising a transparent lid of a plastic Petri dish (150 mm diameter × 10 mm height) (Corning; Corning, NY, USA) and a pale white board was used (Fig. 1A). The arena was illuminated from above by diffuse fluorescent light (350–800 nm; light intensity: 4.22 × 1014 photons/cm2/s). One- to two-day-old flies were transferred in a flask containing a piece of Whatman paper soaked with distilled water, starved for 24 h, and then moved into an empty flask with no water for 6 h before food training. In the following steps, flies were treated individually. Flies were injected into the circular arena through a center hole in the white board by their intrinsic negative geotactic climbing. Loaded flies were trained with a cuboid of the rearing diet (10 × 10 × 5 mm) or a mimic – an odorless food-color sponge the same size as the food – under either the ambient GMF or a modulated MF for 2 min. In each trial, the food or the mimic was placed at one of the four directional locations (i.e., 0°, 90°, 180°, or 270° clockwise relative to the magnetic north for the training), 5 mm from the rim of the Petri dish, similar to a previous study (Dommer et al. 2008). No potential topographic reference was provided for flies to obtain geographic information, e.g., geographic north. The location for the food was wiped with a 70 % ethyl alcohol cotton swab. Flies that failed to continuously remain on the food for 2 min were discarded, and six of the trained flies were prepared for each directional location in a random sequence in each set of experiments. The food-trained flies were transferred to an empty transparent collecting tube for a 2-min rest and were subsequently reloaded into the arena for a 5-min test, under either the same GMF used in the training or a testing MF, depending on the experiment, as indicated in the figure legends. As a control, six of the non-trained flies were rested for 2 min and tested in each set of experiments. The arena was cleansed between trials with a 70 % ethyl alcohol cotton swab. Horizontal movement of tested flies was recorded with a camera (HMX-F90; Samsung, Suwon, Republic of Korea) during the test. Each of the following bearings was determined by averaging the 5-sec interval direction vectors. The direction vector for the “magnetic bearing” of each fly was calculated as the clockwise angle of the fly location with respect to the magnetic north during a test period and the “trained magnetic bearing” was calculated as (clockwise angle for the fly location with respect to the magnetic north during a test period) – (clockwise angle for the food direction with respect to the magnetic north during the fly training) in each test, using a modified method from a previous study (Dommer et al. 2008). Likewise, the “geographic bearing” and the “trained geographic bearing” were calculated as the clockwise angle for the fly direction with respect to the geographic north during a test period and (clockwise angle for the fly location with respect to the geographic north during a test period) – (clockwise angle for the food direction with respect to the geographic north during the fly training), respectively. The trajectory of each fly was analyzed by the video tracking software (ABCLab Co., Daegu, Republic of Korea) used in a previous study (Kim et al. 2016). The experiments were performed double-blind. One experimenter conducted the behavioral assay with a random sample sequence, and another experimenter modulated the GMF conditions and analyzed the data, so that exact information about the samples and the data was not revealed to any experimenter during the experiments or data analysis.
Statistical analysis
To determine the significance of the orientation of flies, circular statistics were performed using Rayleigh's test for each group orientation, and Moore's paired test for paired group comparisons, using Oriana 4 (Kovach Computing Services, Pentraeth, Wales, UK). The statistical parameters were: α, group mean vector as clockwise degree; r, length of group mean vector; S.E.M., standard error of mean. In Rayleigh's test, the mean vectors were flanked by their 95 % confidence interval limits and it was regarded as significant at P < 0.05.
Results
Ambient GMF and eastward-turned MF are used as sensory cues in horizontal orientation
We wondered whether subjecting flies to contrasting declination values would produce declination-dependent locomotive orientation in an open-field arena. Under a food conditioning paradigm, starved flies were individually released into a flat arena to be trained with food under ambient GMF (Table 1), provided with a rest, and then were individually tested without the food in the arena to determine whether they oriented toward the direction of the food relative to magnetic north under the ambient GMF (Dommer et al. 2008) (for details see Materials and methods) (Fig. 1). The food-trained flies significantly orientated toward the direction of the food relative to the magnetic north but the mimic-trained flies did not (Fig. 2A,B and Table S1), and the non-trained naïve flies did not show notable orientation toward any particular direction (Fig. S1A and Table S1). The trajectories of the food-trained flies and the mimic-trained flies were very different as shown in the representative examples (Fig. 1C,D). These results suggested that starved flies trained with food under the ambient GMF could memorize the direction of the food relative to the GMF and use it to forage for food. To confirm whether flies used GMF as a directional reference for horizontal orientation, starved flies were trained under the ambient GMF, and were then tested under an eastward-turned magnetic field (MF) with the same total intensity and inclination but with a 90°-eastward declination (Table S1). Surprisingly, the food-trained flies moved significantly in the direction of the food relative to the magnetic north of the eastward-turned MF (Fig. 2F and Suppl. Table S1) that was ~90° eastward from the food-trained flies' orientation under the ambient GMF above (Fig. 2B) (Moore's paired test: R’ = 2.246, P < 0.001). It is important to note that the flies apparently did not use the geographic direction as a sensory cue in the orientation. Neither the mimic-trained flies nor the naïve flies, however, showed notably orientated movement under the experimental condition (Fig. 2E and Fig. S1B, Table S1). These results indicated that the food-trained flies memorized the direction of the food relative to the magnetic north of the ambient GMF, and then used the eastward-turned MF as a directional reference for the significant horizontal orientation for food searching. Taken together, the results demonstrate that flies can sense the direction of the MF and use it as a directional reference and a sensory cue for horizontal directional movement.
East–west and north–south component-varied declination are used as directional cues
Declination is the angle between magnetic north and geographic north at a certain position on Earth, and is determined based on the relationship declination = arctan (Y/X), where X is the north–south component and Y is the east–west component of the GMF (Finlay et al. 2010). According to the GMF model of IGRF-12, the approximate variation range for declination, X, and Y is −50° – 50°, except around the magnetic poles, −8.77 – 41.53 μT and −17.4 – 9.00 μT, respectively, depending on Earth's surface location (Thébault et al. 2015). Nevertheless, they may not be reliably calculable for a given place using a standard GMF model, such as the international geomagnetic reference field (IGRF), because of local magnetic inhomogeneity (Finlay et al. 2010; Thébault et al. 2015). In addition, the extent of the declination anomaly was found to be up to 28° between areas several kilometers away (NOAA 2016). We sought to further investigate whether declination variation in the natural range could be a sensory cue for horizontal movement. To do this, we adopted three relevant MF conditions (MF 1, 2, and 3) with the same or relatively small differences in either declination, X or Y, that were within the comparable range above, while total intensity and inclination were the same for all conditions (Table 1). First, fruit flies were trained under MF 1 (D = 14.03°, X = 23.69 μT, Y = 5.92 μT) and tested under MF 2 (D = −14.03°, X = 23.69 μT, Y = −5.92 μT). The declination difference is in the inverse sign of Y (east–west component) only. Strikingly, the food-trained flies showed significant movement toward the approximate direction of the food relative to the magnetic north of MF 2, whereas the mimic-trained flies and the naïve flies did not show significant movement in a particular direction (Fig. 3A,B and Fig. S2A, Table S2). These results suggested that the food-trained flies distinguished the relatively small difference in magnetic directions between MF 1 and MF 2, i.e., declination difference for the foraging of food. Another set of fruit flies were trained under MF 2 and then tested under MF 1. Among the three groups of fruit flies, only the food-trained group showed significant movement toward the direction of food relative to the magnetic north of MF 1 (Fig. 3C,D and Fig. S2B, Table S2). The flies were consistent not to use the geographic direction as a sensory cue in the orientation. The difference between the food-oriented direction and that of Figure 3B was approximately 30° (Moore's paired test: R’ = 1.256, P < 0.01), which was comparable to the declination difference between MFs 1 and 2, 28°. Taken together, the results show that fruit flies can sense and use a natural range of declination variance produced by a change in the east–west component of the GMF for horizontal orientation.
The results above prompted us to examine whether differences in X (north–south component), instead of Y (east–west component), could be a sensory cue for horizontal orientation. Flies were trained under MF 2 (D = −14.03°, X = 23.69 μT, Y = −5.92 μT) and tested under MF 3 (D = 14.03°, X = −23.69 μT, Y = −5.92 μT). The food-trained flies showed significant orientation in the approximate direction of the food relative to the magnetic north of MF 3, whereas the mimic-trained flies and the naïve flies did not show notable movement in any particular direction (Fig. 3E,F and Suppl. Fig. S2C, Suppl. Table S2). Next, we subjected another set of flies to MF 3 for training, and then tested the trained flies under MF 2. In line with the results above, the food-trained flies significantly moved toward the approximate direction of the food relative to the magnetic north of MF 2, whereas the mimic-trained flies and the naïve flies did not show any notable orientation (Fig. 3G,H and Suppl. Fig. S2D, Suppl. Table S2). The orientation difference between the food-oriented flies and those in Figure 3F was approximately 26° (Moore's paired test: R’ = 1.218, P < 0.025), which was close to the declination difference between MFs 2 and 3, 28°. Again, the flies did not use the geographic north as a directional reference in the orientation. Taken together, the results demonstrate that fruit flies can use the declination difference generated by variation in the east–west component of the GMF, within a natural range, as a sensory cue for horizontal orientation.
Discussion
The results in the present study show that different magnetic directions due to a natural range of change in either the east–west or north–south component of the GMF can be discriminated and used for horizontal food-search orientation by flies. It has been thought that declination could be determined by magnetoreceptive animals only when they could identify the geographic north using a celestial cue, like a star or sunlight. Based on this assumption, at a glance, it is implausible that the flies in our study used a declination compass without any topographic reference to locate the geographic north in the arena. Nonetheless, contrary to the conventional assumption, our results clearly suggest that declination can be used as a sensory cue to direct the horizontal orientation of flies in a defined physiological context, i.e., starvation, as follows.
Previous studies showed that fruit fly adults and larvae trained for light-attractive and -repulsive orientation, respectively, under ambient GMF or GMF-resembling MFs with different directions, significantly orientated toward the same and opposite direction of light relative to the testing MFs, respectively (Dommer et al. 2008; Phillips & Sayeed 1993). While these studies revealed that fruit flies could use a magnetic compass for horizontal orientation without other helpful directional cues, the results did not show direct evidence for either inclination or any other magnetic compass that was used for the orientations. In addition, it was not possible to examine whether a natural range of declination change caused by altering the east–west or north–south component of the GMF alone influenced on the orientation in these studies, since the experimental setup did not include such conditions to test (Phillips 1986). Therefore, it is reasonable to consider that there has been no convincing evidence for the magnetic compass that underlies the observed flies' horizontal orientations. Second, recent studies that showed magnetosensitive body alignment in two animal species (Bazalova et al. 2016; Hart et al. 2013) supported the possibility that animals can exhibit magnetoreceptive behavior by sensing changes in declination without detection of the direction of geographic north. Cockroaches could sense repeated changes in magnetic direction caused by the shift of the horizontal component of the GMF; they showed significant correlated body alignment within an interval of a few minutes without any geographic directional cue in the testing arena (Bazalova et al. 2016). In addition, the north–south body alignment of domesticated dogs during excretion in open fields was specifically dependent on the rate of change in the declination, but not the absolute value of declination or total-, vertical-, horizontal-intensity of the GMF (Hart et al. 2013). Thus, very subtle change in declination of the order of 0.0007°/min was enough to disturb the significant body alignment, highlighting the high sensitivity of these animals to declination change. In particular, the north–south body alignment did not depend on the time of day, presence of the visible sun, or the routes of walks, indicating that the body alignment did not require apparent celestial or topographic cues for geographic direction, which has been assumed to be necessary for sensing declination in magnetoreceptive animals. Third, the declination compass rather than the inclination compass may be more persuasive in explaining the magnetic direction-dependent orientation of the fruit flies in the present study. Magnetoreceptive animals such as birds, sea turtles, and butterflies, showed equatorward or poleward directional movements that were parallel to the magnetic field lines using the inclination compass derived from the radical pair mechanism (Guerra et al. 2014; Wiltschko & Wiltschko 2006). Fruit flies have also been shown to use the cryptochrome-mediated radical pair mechanism for various magnetosensitive behaviors (Bae et al. 2016; Fedele et al. 2014; Gegear et al. 2008). If the inclination compass was used for orientation, the food-trained flies would have oriented themselves toward the magnetic north or magnetic south of the testing MFs irrespective of food direction relative to the MFs during the training, which might have led to a magnetic north–south axis orientation in the tests. How did we then envision that the flies could determine and use declination as a directional cue? The daily variation of declination reaches a few tenths of a degree even in the absence of solar disturbance (Finlay et al. 2010; Thébault et al. 2015). Although the scale of variation appeared to be very small, magnetosensitive animals such as dogs appeared to be able to detect such minute variation as described above (Hart et al. 2013). If this also somehow happened in the fruit flies, they were probably able to use both the inclination and declination compasses at the same time. For instance, flies could detect the magnetic north–south axis with enough fidelity using the inclination compass despite the minute variation in declination. Further, when the flies detected fluctuations in the magnetic direction, i.e., an eastward or westward shift relative to the magnetic axis, they possibly considered the shifting direction as the east–west axis, which was their declination cue. Once the flies were able to determine both directions for the axes, they could use the two compasses together to search for food during the food training and the test.
It is likely that flies have evolved to incorporate a declination compass into their multi-modal sensorimotor system. In terms of compass modality, declination could be complimentary to or more advantageous than inclination because it can confer more sensitive directional cues especially for short- or middle-distance migration. In addition, the ability to use declination, a much more consistent environmental cue compared to other sensory stimuli such as light, odor, and sound, might have been highly advantageous for securing a more efficient toolset for foraging or migration. A very recent related theoretical study suggested that GMF-exploiting migratory birds probably need another navigation system, e.g., an olfactory map to compensate for the variable fidelity of returning sites of the magnetic map, which is based on GMF intensity and inclination (Komolkin et al. 2017). Nonetheless, local olfactory cues are highly dependent on temporal and spatial weather conditions, which considerably reduce the feasibility of olfactory maps in nature. If declination coordinates rather than the olfactory map were integrated into the GMF map, the tri-modal magnetic map would presumably increase site fidelity and exhibit higher consistency than the GMF map that is based on the total intensity and inclination alone. However, daily and seasonal variations of declination are often remarkably augmented by disturbances from solar storms, earthquakes, and local anomalies such as mountains, rocks, and even various artificial structures. Such unpredictable variations within a certain range may provide a stimulus for flies to disperse from old habitats to new places, which could be a hitherto unknown cause and mechanism for gene flow among fruit flies.
Further study is required to address some important questions, such as how the flies perceive different directions with respect to the same geomagnetic direction, and thus sense the north–south and east–west components of the GMF to discriminate different declinations. Finally, it may be necessary to reexamine whether declination influences the magnetosensitive behaviors reported in many previous studies, because our calculation using the series of IGRF models showed that the unclarified declination values as well as the values of total intensity or inclination in most of previous magnetoreceptive studies varied substantially.
Acknowledgments
We thank Dr. Peter Hore (University of Oxford, UK) for his valuable comments on the draft manuscript. This research was supported by the Basic Science Research Program (2015R1D1A3A01019256) and the Basic Research Laboratory (BRL) Program (2013R1A4A1069507) to KSC through the National Research Foundation (NRF) funded by the Korean Ministry of Education. The authors report no conflicts of interest.
