Odor plumes and animal navigation in turbulent water flow: a field study.
We experimentally establish a direct link between the transport of chemical stimuli and animal navigation in a natural, turbulent flow environment. Attraction of predatory blue crabs (Callinectes sapidus) to odor released by clam prey (Mercenaria mercenaria) was investigated. The mechanisms directing predator to prey were identified for crabs foraging naturally in estuarine tidal creeks, where water flows unidirectionally for hours at a time. These creeks were shallow enough to permit direct, noninvasive observations of crab locomotory behavior.
Experiments were conducted in the North Inlet Estuary, near Georgetown, South Carolina, USA (32 [degrees] 20[minutes] N, 79 [degrees] 15[minutes] W). Flow records were obtained with an electro-magnetic flow meter (Marsh-McBirney 523) equipped with a two-dimensional sensor (1 cm diam), mounted on a flat base, and submerged in the tidal creek. The base was positioned flush with the sandy bottom, and the sensor was placed 5 cm above the substrate. Sensor height was chosen to match the elevation of a typical, adult crab body. A data logger (Campbell CR10) was used to record both horizontal and vertical flow velocities continuously measured by the sensor at 1 Hz, between 26 June and 8 August 1993. Horizontal flow (downstream advection) typically ranged from 0 cm/s at slack tides, to 30 cm/s during peak ebb and flood tides. We applied the eddy correlation method to calculate shear velocities, at 1-min intervals, using the correlation between the horizontal and vertical flow velocities (8, 9). Shear velocity ([u.sub.*]) is a measure of the strength and correlation of turbulent fluctuations in flow speed near the substratum. Finally, we determined the coefficients of turbulent mixing as the products of shear velocity, sensor height above the substratum, and Von Karman's constant (9). These mixing coefficients were remarkably low, ranging from 0.5 to 1.5 [cm.sup.2]/s. They indicated that across-stream mixing occurred very slowly in the tidal creek even though water flow was turbulent.
The dynamics of odor plumes were characterized by measuring the transport of fluorescent dye (sodium fluorescein) and an electrochemical (dopamine) following their combined release from a point source. Input concentrations of fluorescein and dopamine were 1.0 g/liter and 2 mmol/l, respectively, with ascorbic acid added to the mixture at 20 mmol/l as an antioxidant. The mixture was introduced through polyethylene tubing (0.5 mm ID) at 6 ml/min.
Fluorescein provided a visual marker, and fluorometric determinations were used to establish the time-averaged distributions of dye at downstream and at across-stream sites, relative to the release point. Water samples were collected (at 1 ml/s) simultaneously over 1-min intervals by syringes placed at each of 18 to 30 sites per trial. These sites were distributed in a grid, with six sites placed across-stream (0, 2, 4, 6, 8, 10 cm distant from the plume mid-line) at three to five locations downstream of the release point (5, 25, 50, and in some trials, 100 and 275 cm distant from the source). The bell-shaped distribution of fluorescein concentration with gradual decay downstream [ILLUSTRATION FOR FIGURE 1 OMITTED] is what a Gaussian plume model would predict (Pearson's product-moment correlation: [r.sup.2] [greater than or equal to] 0.95; P [less than] 0.001; all replicate plume measurements). Concentration dropped sharply at the plume's visible lateral edges [ILLUSTRATION FOR FIGURE 1 OMITTED].
Fluorometric measurements also provided an alternative method of calculating the mixing coefficient for comparison with determinations made using the electro-magnetic flow meter. Temporal changes in the across-stream variance in fluorescein concentration were used to estimate the mixing coefficient (9). Results from the two methods matched well. For example, estimates based on fluorometric determinations ranged from 0.5 to 1.2 [cm.sup.2]/s during a time when estimates of 0.5 to 0.8 [cm.sup.2]/s were made from electromagnetic flow meter records.
When measured at fast temporal scales, chemical stimuli in odor plumes are patchily distributed due to turbulence. Mean concentrations and time-averaged distributions of fluorescein dye, therefore, may not be indicative of the information available for crabs attempting to orient towards an odor source (3, 4, 5, 6, 7). Because arthropod chemoreceptors detect intermittent (or pulsed) chemical stimuli applied at a maximum frequency of 4-10 Hz (10, 11), we employed carbon fiber microelectrodes (150 [[micro]meter] diam) and a computer recording system (MedSystems Corp. IVEC-10) to sample dopamine at 10 Hz (12). Electrode recordings were made at the fluorescein sampling sites (see above). Turbulent mixing caused the concentration of dopamine sampled downstream of the source to fluctuate strongly in time and space. Bursts of highly concentrated chemical passed over the sensor, alternating with periods of low or zero concentration [ILLUSTRATION FOR FIGURE 1 OMITTED]. The plume's lateral edge, as defined by our high-speed dopa-mine measurements, was positioned identically relative to the edge detected both by our time-averaged fluorescein measurements and by our visual observations. This lateral edge, separating clean from chemical-laden water, was very narrow (2-4 cm wide) compared to the body size of an adult crab (10-15 cm carapace width). Thus a steep concentration gradient was found across-stream, but not downstream of chemical release.
We previously demonstrated that some crabs search for and find intact live clams, and that these crabs are responding to odor plumes created by the excurrent release of attractant metabolites at low concentration (13). Once a clam is found, however, it is chipped open by a crab and attractants are released to form a plume of high concentration. High-concentration plumes then immediately attract other crabs to the predation site. Hence, depending on the situation, crabs may be exposed either to low or high attractant concentrations, and crabs respond effectively in each case.
Concurrently with hydrodynamic and chemical measurements, we assessed crab orientation in odor plumes. Our field studies focused on plumes characteristic of chipped clams. We chose to work with chipped clams because high attractant concentrations would better ensure an effective stimulus through the broad range of hydrodynamic conditions encountered by crabs in the field. Stimulus plumes (dyed with fluorescein for visibility) were created, either presenting a chipped clam or introducing clam mantle fluid (membrane filtered to 0.22 [[micro]meter]) at a rate mimicking its release from a chipped clam. Each stimulus plume was always paired with a control plume that delivered fluorescein in filtered seawater. Both stimulus and control solutions were introduced at 6 ml/min, with inputs separated by 60 cm across-stream.
Free amino acid compositions of effluent leaking from chipped clams (n = 8 clams assayed), mantle fluid of intact live clams (n = 8 clams), and homogenized clam flesh (n = 7 clams) were all determined using a Beckman 6300 System Gold high-performance liquid chromatograph with a sodium ion-exchange column (4-mm ID x 120 mm; Beckman) for separation. In this system, amino acids were monitored spectrophotometrically after post-column reaction with ninhydrin. Compositions of clam effluent and mantle fluid were almost identical (Pearson's product-moment correlation: [r.sup.2] = 0.998; P [less than] 0.001; n = 18 amino acids chromatographed), indicating that mantle fluid was the source of effluent material. Because taurine was by far the most abundant amino acid in both clam effluent and mantle fluid (accounting for [greater than]50% of the total amino acid composition), we used it as a marker to measure the rates of fluid release from chipped clams. In the laboratory, clams (n = 12) of various sizes were chipped by using a metal rod to deliver a single firm blow to the lateral shell margin. The resulting chip was similar in size and shape to one produced by a blue crab as it begins to feed. Each chipped clam was then placed individually into a separate beaker of artificial sea-water. The beakers were stirred, and they were maintained at the same temperature and salinity as seawater in the tidal creek from which clams were collected. Artificial seawater was sampled from each beaker before placement of the clam, and again at 30-to 60-s intervals for 15 min after placement; HPLC analysis of this seawater indicated that taurine (and mantle fluid) release was constant throughout the trial period. The relation between taurine release rate and clam size was then used to scale our delivery of mantle fluid in field experiments, simulating the fluid input from an intermediate-sized chipped clam (6 cm total shell length).
Quantitative observations of foraging crabs were made noninvasively with a video camera (Sony TR81) mounted 4 m above the tidal creek. Video records of crabs responding to plumes were made during ebbing tides. A 3-m field of view was dictated by the resolution of the video camera and the size of the crabs (10-15 cm carapace width), whose positions could not be reliably quantified in wider field images. A scale bar in the field of view was employed to measure distance and to correct for distortion due to perspective. In the laboratory, plume edges and crab positions at 1-s intervals were traced onto acetate sheets from video playbacks to a monitor. Both crab location, in relation to the plume edge, and crab locomotory kinematics were measured [ILLUSTRATION FOR FIGURE 2 OMITTED].
Crab responses to the plumes were dramatic and unambiguous: 29 crabs contacted the control plumes, but only 4 of these crabs walked upstream towards the input source. The near absence of positive responses to control plumes demonstrated a lack of attractant effect by fluorescein dye. In comparison, after contacting stimulus plumes, 68 of 80 crabs walked upstream to the input source. Crabs turned upstream within 1.5 s ([+ or -]0.3 SD) of contacting an odor plume. Percentages of crabs responding positively to plumes from either chipped clams or mantle fluid were nearly identical, being 86% (n = 52 crabs) or 82% (n = 28 crabs), respectively (G-test for homogeneity: [G.sup.2] = 0.270; df = 1; P [greater than] 0.50). Our results demonstrate an odor-conditioned rheotaxis that orients crabs upstream. Previously, we reached an identical conclusion for blue crabs foraging in a laboratory flume. Crabs walked upstream to find intact live clams in flowing water, but they oriented indiscriminately and searched unsuccessfully for clams in still water (13).
Oriented movements by crabs lateral to water flow are controlled by chemotaxis. As crabs walked upstream towards an attractant source, they frequently approached the lateral edges of the plume. When crabs did reach the edge, they nearly always turned directly back to the plume (50 of 61 turns; G-test for goodness-of-fit, 1:1 hypothesized ratio, [G.sup.2] = 14.97; df = 1; P [less than] 0.001), without exhibiting either casting or zigzagging [ILLUSTRATION FOR FIGURE 2 OMITTED]. Lateral movements were initiated as crabs began to exit a plume and partially contacted clean water. Fluorescein did not act as a visual cue, because crabs displayed identical oriented responses when tests were conducted in the dark (under infrared illumination) and without fluorescein (13, and in prep.). It took, on average, less than 1 s (0.8 [+ or -] 0.2 s SD) for crabs to renew upstream walking after they had begun moving laterally towards the plume midline. Remarkably, we did not observe walking speeds to change significantly as crabs moved closer to attractant sources (analysis of covariance: F = 0.60; df = 4,237; P = 0.66; walking speed: 12.8 [+ or -] 0.4 cm/s SD), and we found no significant correlation between walking speed and water flow velocity (Pearson's product-moment correlation: [r.sup.2] = 0.037; df = 1,64; P = 0.12). We hypothesize that crabs perceive clam attractant as a binary cue (present/absent), both in their upstream movement and in their across-stream walking. Because the plume edges were very sharp, when crabs partially exited the plume, some pereiopods (legs or claws) were outside the plume while others remained inside. A comparison of simultaneous chemosensory inputs from the appendages inside and outside the plume would presumably allow the crabs to determine the correct direction and return to the plume. This binary response would lead crabs to locations of higher concentration of clam attractant.
Orientation mechanisms used by crabs in upstream movement are similar to those of flying insects. However, crabs differ from insects in their across-stream response. Insects provide the only other system in which navigation relative to odor plumes has been coupled with fluid dynamics. Flying insects locate a source of chemical attractant by moving upwind upon contacting a filamentous trace of attractant odor (14, 15). After several seconds of flying in clean air, insects shift to casting (regular reversals of flight directed across-stream) until contact with another odor trace causes a return to upwind flight (16). Flying insects, therefore, do not use chemical concentration gradients to determine either their upwind or across-wind directions (16, 17, 18). The use of chemotaxis may be impractical in their environment, where complex fluid dynamics do not permit stable zones of high attractant concentrations to exist. Crabs in contrast, consistently turn back into the attractant plume rather than zigzagging after losing the plume signal.
The difference between estuarine tidal creek flow and atmospheric winds may explain why blue crabs and insects use contrasting mechanisms for successful navigation towards an odor source. The crop fields and forests used as experimental models for insect flight are hydraulically rough, with high advection. Eddy diffusivities in insect habitats are 100 to 1000 times greater than those we recorded in estuarine tidal creeks (19, 20). Higher diffusivities yield plumes consisting of tiny, highly dispersed filaments or parcels of odor. Wind direction changes frequently, causing plumes to meander (3, 6). The dispersal pattern of odor, coupled with the relatively fast flight speed of insects, means that a flying insect has little chance to detect more than the occasional pulse of passing odor. Casting, zigzagging, and rapid behavioral modulation in response to fine-scale changes in odor concentration may be strategies appropriate for situations in which the entire plume meanders away from the animal.
In contrast, the flow environment of estuarine tidal creeks is markedly less turbulent, yielding relatively stable, straight, and sharply delineated odor plumes. Plumes cannot meander substantially, because flow is constrained by water depth and by the sides of the creeks. A stable plume structure permits direct binary comparisons of chemical concentration inside and outside the plume, to guide movement lateral to flow. The more direct plume-following behavior and across-stream chemotactic responses shown by crabs reflect a strategy appropriate to the plume structure characteristic of their environment. Mechanisms of plume-following behavior, therefore, arise in response to chemical stimulus distributions, as determined by the specific fluid dynamic environments in which animals must naturally navigate.
We thank Dr. D. M. Allen, Director of the USC Baruch Field Laboratory, for providing laboratory space and logistical support for field work. Dr. Y. Ishikawa, USC Institute for Biological Research and Technology, performed amino acid composition analyses. We also thank Dr. S. A. Woodin, whose comments on earlier drafts greatly improved this manuscript. This research was sponsored by the National Science Foundation (IBN 92-22225) and the University of South Carolina Research and Productive Scholarship Fund.
1. Wright, R. H. 1958. The olfactory guidance of flying insects. Can. Entomol. 90:81-89.
2. Bossert, W. H., and E. O. Wilson. 1963. The analysis of olfactory communication among animals. J. Theor. Biol. 5: 443-469.
3. Murlis, J., and C. D. Jones. 1981. Fine-scale structure of odour plumes in relation to insect orientation to distant pheromone and other attractant sources. Physiol. Entomol. 6: 71-86.
4. Elkinton, J. S., R. T. Carde, and C. J. Mason. 1984. Evaluation of time-average dispersion models for estimating pheromone concentration in a deciduous forest. J. Chem. Ecol. 10: 1081-1108.
5. Zimmer-Faust, R. K., J. M. Stanfill, and S. B. Collard, III. 1988. A fast, multi-channel fluorometer for investigating aquatic chemoreception and odor trails. Limnol. Oceanogr. 33: 1586-1595.
6. Murlis, J., J. S. Elkinton, and R. T. Carde. 1992. Odor plumes and how insects use them. Annu. Rev. Entomol. 37: 505-532.
7. Moore, P. A., and J. Atema. 1991. Spatial information in the three-dimensional fine structure of an aquatic odor plume. Biol. Bull. 181: 408-418.
8. Schlichting, H. 1979. Boundary Layer Theory. McGraw-Hill, New York. 486 pp.
9. Denny, M. W. 1988. Biology and Mechanics of the Wave-Swept Environment. Princeton University Press, Princeton, NJ. 329 pp.
10. Kaissling, K. E., C. Z. Straussfeld, and E. Rumbo. 1987. Adaptation processes in insect olfactory receptors: mechanisms and behavioral significance. Annals N.Y. Acad Sci. 510: 104-112.
11. Gomez, G., R. Voigt, and J. Atema. 1994. Frequency filter properties of lobster chemoreceptor cells determined with high-resolution stimulus measurement. J. Comp. Physiol. A 174: 803-811.
12. Moore, P. A., G. A. Gerhardt, and J. Atema. 1989. High resolution spatio-temporal analysis of aquatic chemical signals using micro-electrochemical electrodes. Chem. Senses 14: 829-840.
13. Weissburg, M. J., and R. K. Zimmer-Faust. 1993. Life and death in moving fluids: hydrodynamic effects on chemosensory-mediated predation. Ecology 74: 1428-1443.
14. Mafra-Neto, A., and R. T. Carde. 1994. Fine-scale structure of pheromone plumes modulates upwind orientation of flying moths. Nature 369: 142-144.
15. Vickers, N. J., and T. C. Baker. 1992. Male Helothis virescens maintain upwind flight in response to experimentally pulsed filaments of their sex pheromone (Lepidoptera: Nocturidae). J. Insect Behav. 5: 669-687.
16. Baker, T. C. 1986. Pheromone-modulated movements of flying moths. Pp. 39-48, in Mechanisms of Insect Olfaction, T. L. Payne, M. C. Birch, and C. E. Kennedy, eds. Clarendon Press, Oxford.
17. David, C. T., J. S. Kennedy, and A. R. Ludlow. 1983. Finding of a sex pheromone source by gypsy moths released in the field. Nature 303: 804-806.
18. Arbas, E. A., M. A. Willis, and R. Kanazaki. 1993. Organization of goal-oriented locomotion: pheromone-modulated flight behavior of moths. Pp. 159-198, in Biological Neural Networks in Invertebrate Neuroethology and Robotics, R. D. Beer, R. E. Ritzmann, and T. McKenna, eds. Academic Press, New York.
19. Shaw, R. H., J. Tavangar, and D. P. Ward. 1983. Structure of the Reynolds stress in a canopy layer. J. Climate Appl. Meteorol. 22: 1922-1931.
20. Arya, S. P. 1988. Introduction to Micrometeorology. Academic Press, San Diego. 303 pp.