QUICK SEARCH:   [advanced] Author: Keyword(s):
Home     Help     Feedback     Subscriptions     Archive     Search     Table of Contents

First published online July 20, 2006
Journal of Experimental Biology 209, 3001-3017 (2006)
Published by The Company of Biologists 2006
doi: 10.1242/jeb.02305

This Article Summary Full Text Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Budick, S. A. Articles by Dickinson, M. H. Search for Related Content PubMed PubMed Citation Articles by Budick, S. A. Articles by Dickinson, M. H. Social Bookmarking                         What's this?

Free-flight responses of Drosophila melanogaster to attractive odors

Seth A. Budick^1,* and Michael H. Dickinson^1,2

¹ Division of Biology, California Institute of Technology, Pasadena, CA 91125, USA
² Division of Bioengineering Option, California Institute of Technology, Pasadena, CA 91125, USA

View larger version (36K): [in a new window]   Fig. 1. (A) Schematic profile view of the full wind tunnel. Gray/dotted line at upwind end (right) represents the 76 cm cardboard box used for generation of the homogeneous cloud. Smoke visualization of the ribbon (B) and large diameter (C) plumes. (D) Front view of the 76 cm cardboard box with four mixing fans. (E) Schematic representation of the working section of the wind tunnel. Flies were introduced via a pipette tip glued to the end of a tube at the downwind end of the tunnel; odor was introduced via a second tube at the upwind end. The floor of the tunnel was painted black and the walls were covered with a random checkerboard pattern. Arrays of IR diodes illuminated the tunnel, which was visualized by two IR-sensitive cameras positioned above it (camera positions not to scale). (F) A sample trajectory in a no odor, control treatment. (G) 386 overlaid trajectory fragments from the same treatment. (H) A transit probability histogram derived from the trajectories in G. Transit histograms were derived by dividing two-dimensional views of the wind tunnel into 7220 squares with side lengths of 0.8 cm. Within a given treatment, the number of fly occurrences within each square was summed and divided by the total number of fly occurrences in all squares to yield a probability of square occupancy, where the total probability summed to 1.0. Scale bars: 1 cm in B,C.

View larger version (20K): [in a new window]   Fig. 2. Trajectory parameters subjected to analysis (explanations in text).

View larger version (46K): [in a new window]   Fig. 3. D. melanogaster are anemotactic and manifest a centering response. Transit histograms, viewed from above, of flies released at the downstream end of the tunnel in still air (A, 132 flies) and in the presence of a 0.4 m s-1 wind (B, 80 flies). Based on a mixture of von Mises distributions, flight headings were bimodally distributed up (-2.91±28.30°) and down (177.72±32.45°) the longitudinal axis of the tunnel in still air (C), but were unimodally centered on 1.46±13.67° in the presence of wind (D). Raw counts of instantaneous heading values are plotted in C and D, but all statistical analyses are on mean trajectory headings, which were significantly more dispersed in the absence of wind (N=80, U=2730, P<0.0001). (E) A histogram of the distribution of flies across the tunnel's width indicates that flies manifested a centering response within the central 0.5 m of the tunnel in a 0.4 m s-1 wind. Flight in only the central section was analyzed in order to minimize the visual effects of the odor- and fly-releasing tubes.

View larger version (56K): [in a new window]   Fig. 4. Flies orient towards a conspicuous visual object, a black post (white dot), in the absence, but not in the presence of wind. Transit histograms, with flies viewed from above, in no wind (A, 67 flies) and in a 0.4 m s-1 wind (B, 39 flies). The white circle represents the distance from which the post would subtend 5° on a fly's retina. Without wind, 26% of total fly transit within the tunnel was located within the circle, but with wind this percentage dropped to 11%.

View larger version (18K): [in a new window]   Fig. 5. (A-H) Eight examples of odor-mediated upwind flight with air speed encoded by color. The odor of fermented banana was presented to the flies in the form of a ribbon (1 cm diameter) plume. Flight trajectories, viewed from above, showed substantial variability following plume contact (plume contact indicated by arrow and black dot). Several features are often salient, however, including a shift from cross-wind to upwind flight as well as a fast upwind surge. Upwind progress is often interrupted by looping counterturns and casting flight directed across wind.

View larger version (59K): [in a new window]   Fig. 6. Flies localize a ribbon plume of banana odor. Trajectories were partitioned into `pre-contact' and `post-contact' fragments. Transit histograms of post-contact flight indicate that flies localized and maintained close proximity to the plume (white bar) both in the horizontal (A) and vertical (B) dimensions (278 episodes of plume contact from 127 flies). In the absence of an odor plume, flies distributed much more uniformly throughout the tunnel (C and D, horizontal and vertical dimensions respectively, 51 episodes of `plume contact' from 36 flies). Note that the plume did descend slightly along the tunnel's length and that the white bar accurately represents the approximate plume extent.

View larger version (29K): [in a new window]   Fig. 7. Following plume contact, flies fly faster, straighter upwind. (A) Prior to plume contact, flight headings were distributed trimodally, with modes at 0.00±18.16°, 84.77±49.35° and -80.89±41.79°, based on the fit of a mixture of three von Mises distributions to the raw counts of instantaneous heading vectors. The shaded curve represents the trimodal model fit. (B) Following plume contact, flight was unimodally directed upwind (2.18±55.92°). For statistical analysis, mean pre- and post-contact headings were calculated for each fly. Mean pre-contact headings were significantly more dispersed than the corresponding post-contact means (N=138, U=5507, P<0.0001). (C) Proportions of the total counts of instantaneous trajectory values for upwind velocity, air speed and plume distance. Comparing trajectory means for each fly, upwind velocity increased following plume contact (t=4.53, d.f.=223.11, P<0.0001) as did air speed (t=3.71, d.f.=250.94, P<0.001), while the flies remained closer to the plume, (t=5.46, d.f.=273.31, P<0.0001) (pre-contact, empty bars; post-contact, filled bars, 138 flies).

View larger version (36K): [in a new window]   Fig. 8. Plume contact results in trajectories oriented along the plume line. 35 randomly selected episodes of plume contact were translated and aligned at the point corresponding to entrance into the plume cylinder in the banana odor ribbon plume and in a no odor control. (A) In the banana odor plume, pre-contact flight was largely directed across-wind whereas plume contact was followed by a shift towards upwind progress while maintaining close proximity to the plume-line with flies occasionally casting across wind. (B) In the absence of an odor plume, `plume' contact lacked consistently similar effects on trajectory structure.

View larger version (27K): [in a new window]   Fig. 9. Effects of contact with a ribbon plume of banana odor on kinematic parameters. The mean and standard error envelope are plotted for 1 s of flight prior to and following the first episode of plume contact in the banana odor ribbon plume (yellow error envelope, 124 flies) and in a no odor control (gray error envelope, 44 flies). Because not all trajectories consisted of at least 1 s of flight prior to and following plume contact, means and standard errors were calculated at all time points from all trajectories whose durations met or exceeded that threshold length.

View larger version (26K): [in a new window]   Fig. 10. In a homogeneous cloud, flight is directed almost completely upwind. (A) In a no odor control, wind polarized flight upwind (0.49±21.98°, 80 flies), but still resulted in greater dispersion of mean heading angles (N=80, U=2995, P<0.05) than in a homogeneous odor cloud (B, -1.16±31.16°, 94 flies). Note that raw heading counts are plotted, but only mean trajectory headings are analyzed statistically. Upwind flight was also significantly faster in the homogeneous cloud (d.f.=171.84, t=4.57, P<0.0001) (C, no odor, empty bars; homogeneous cloud, filled bars).

View larger version (18K): [in a new window]   Fig. 11. Representative trajectories illustrate the differences between flight in a homogeneous cloud, clean air, and a banana odor ribbon plume. Four representative, `post-contact,' trajectories are shown from (A) a homogeneous cloud, (B) clean air and (C) a banana odor ribbon plume. Flight in the homogeneous cloud often gave rise to very straight upwind trajectories compared to clean air. Trajectories in clean air headed generally upwind, while those in the banana odor ribbon plume were largely characterized by upwind flight interspersed with cross-wind casts.

View larger version (16K): [in a new window]   Fig. 12. Casting frequently follows plume truncation. (A) Four representative trajectories from the large diameter pulsed banana odor plume illustrate flight prior to plume contact (red), within the plume (blue), and following plume loss due to truncation (gray). The first cast, as defined by our cast identification algorithm, is plotted in green. (B) In the continuous large diameter plume, casting rarely initiated within the plume (color designations as above except that gray indicates plume loss due to flight out of the plume rather than plume truncation). Arrows indicate the initiation of fly tracking, but in some cases several points were excised from the beginning of the track in order to enhance the clarity of the trajectory.

View larger version (15K): [in a new window]   Fig. 13. Plume loss increases the probability of casting. (A) To test the effect of plume truncation on the probability of cast initiation, we calculated the mean duration of contact with a pulsed plume prior to truncation (383 ms). (B) The probability of casting following plume truncation was compared to that following 383 ms exposure to the continuous plume. The probability of cast initiation within each 50 ms bin following plume truncation, or `pseudo-plume truncation', was then calculated as described in the text. Casting was significantly more likely following plume truncation than in the continuous plume (d.f.=1, 2=8.96, P<0.01), with 29.6% of flies initiating a cast within 1 s of plume truncation with a mean latency of 330±140 ms (Di). For flies in the pulsed plume, cast initiation was significantly more likely following plume truncation (Di) than following plume contact (Diii; d.f.=1, 2=6.66, P<0.01).

View larger version (31K): [in a new window]   Fig. 14. (A-F) Effects of cast initiation on trajectory parameters. Kinematic parameters were aligned at the moment of cast initiation following plume loss due to truncation (mean and envelope of standard deviation are plotted). Casts manifested velocity profiles that were nearly the inverse of the plume contact response, with turns clearly initiating prior to reaching the 50° heading threshold necessary for the cast identification algorithm (C, 24 casts).

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

Twitter