Table Of Content

Reference: Bio/. Bull. 198: 203-212. (April 2000) Mechanisms of Animal Navigation in Odor Plumes NEIL VICKERS J. Department ofBiology, University ofUtah, Salt Lake C/rv. Utah 84112 Abstract. Chemical signals mediate many of life's pro- mechanical conditions in the environments that these dif- cesses. For organisms that use these signals to orient and ferent macroscopic organisms inhabit. navigate in their environment, where and when these cues are encountered is crucial in determining behavioral re- Introduction sponses. In air and water, fluid mechanics impinge directly upon the distribution of odorous molecules in time and By and large, life exists in one oftwo fluidenvironments, space. Animals frequently employ behavioral mechanisms air or water. These media offer a vast array of different that allow themtotake advantage ofboth chemical and fluid conditions, even the most severe of which have been suc- dynamic information in order to move toward the source. In cessfully colonized by life. Although the range of condi- turbulent plumes, where odor is patchily distributed, ani- tions that support life appearstobeextremelybroad, thefact mals are exposed to a highly intermittent signal. The most thatairand waterare fluids meansthat, in principle, they are detailed studies that have attempted to measure fluid dy- governed by the same physical laws (Denny, 1993). Thus namic conditions, odor plume structure, and resultant ori- the physical behavior of fluids is important to organisms entation behavior have involved moths, crabs, and lobsters. that orient with respect to chemical signals, because fluid The behavioral mechanisms employed by these organisms dynamics impinge directly upon the distribution ofodorant are different but generally integrate some form of chemi- molecules in time and space. Furthermore, the fluid envi- cally modulated orientation (chemotaxis) with a visual or ronment also impinges directly upon the locomotory activ- mechanical assessment of flow conditions in order to steer ity of organisms. Hence, fluid conditions dictate both the up-current or upwind (rheo- or anemo-taxis, respectively). dispersal of odor molecules in the environment and the Across-stream turns are another conspicuous feature of abilityoforganisms toorienttothe signal beingtransported. odor-modulated tracks ofa variety oforganisms in different The physical nature of air and water with particular refer- fluid conditions. In some cases, turning is initiated by de- ence to biological systems has been excellently reviewed in tection ofthe lateral edges ofa well-defined plume (crabs), considerable detail elsewhere (Denny, 1993; Vogel, 1994; whereas in other animals turning appears to be steered Weissburg, 1997, 2000), so the brief discourse here will according to an internally generated program modulated by serve as the framework for the more substantive treatment odor contacts (moth counter-turning). Other organisms such that will be given to the mechanisms employed by animals as birds and fish may use similar mechanisms, but the to navigate and orient with respect to chemical signals. The experimental dataforthese organisms is notyet as convinc- focuswillbe on macroscopic animals, althoughexamplesof ianngi.maTlhserebseuhltaviinoroarliensttartaitoegniersesepomnpsleosyethdatbayreaapvparrioeptryiatoef smaTlhleerabliilfiet-yfotormssenwsilelabnedurseesdpownhdetroetaheppcrhopermiiactael.s present in the environment exists in virtually all living creatures for the dispersed, intermittent plumes dictated by the fluid- from unicellular bacteria to complex, multicellular organisms. The scale of these organisms covers many orders of Received 6 July 1999; accepted 24 November 1999. magnitude, from the smallest single cell to the largest crea- E-mail: [email protected] tures that exist today. Biological as well as physical limita- This paper was originally presented at a symposium titled Chemical tions must, therefore, have shaped the types of chemical DCioemgmou,niCcaalitfioornniaa,ndonEc3o0logDye.ceTmhbeersy1m9p9o8s.iuwma,s wihnviicthedwbays thheeldWeisntSearnn signals that a particular organism can detect and respond to Society ofNaturalists as part ofits annual meetings. effectively. 203 204 N. J. VICKERS What is an Odor I'!mm. or How do Fluid I'lnsics the tact that fluid flow is rarely stable. Instead, swirling and Affect tht- Distribution of Odor Molecules eddies within the moving air seem to play a major role in in Time tind Space? distributing smoke particles. The chaotic movements ofthe fluid are known as turbulence and are caused by fluctuations When molecules escape a solid substrate the) become in velocity. Turbulence may be caused by. among other subject to physical forces in the fluid. \\'hat features of things, the flow offluids around stationary objects (that may fluids are important in determining how molecules are dis- or may not be acting as sources ofchemical emission) that persed once those molecules become fluid-borne'.' In a fluid induce the formation of eddies and vortices in their wake. that is not moving, molecules that escape the substrate form Various mathematical terms can be used to describe the a concentration gradient that is highesi closest to the source nature ofthe fluid environment (c.i;.. the Reynolds number). and lowestat points farthestaway. Oh\louslv. the molecular A complete description ofthese terms and their implications weight of a molecule has a significant impact on the time for fluid flow can he found in extensive reviews elsewheie taken foraconcentration gradient in be established through- (Denny. 199V Vogel. 1994: Belanger and \\illis. 1996: out a given \olume of fluid \\ here molecular diffusion is Weissburg. 1997. 2000). the main force driving the distribution of molecules, the Importantly, particles being transported in turbulent ed- only signal available to .mial is the chemical itself and dies have a tendency to move down a concentration gradi- its spatiotemporal gr.ul' ent, but turbulent diffusion is a property of how the fluid For organisms oi om scale it seems hard to envisage moves and not of the particles themselves. Therefore, in a situations in which fluids arc complete!) static and molec- plume created bv turbulent diffusion, unlike one produced ular diffusion is the predominant force responsible for dis- by simple molecular diffusion, the molecular weight of a tributing molecules. However, many organisms exist at molecule does not impact its distribution. Consequently, scales where molecular diffusion is the salient force behind molecules that are released from the substrate at a certain the distribution of chemical signals. Take a bacterium, for ratio should retain their relative concentrations in situations example, that exists within a small volume of fluid. Diffu- where turbulent (low dominates. Hence, odor blends could sion will distribute molecules emitted h\ a source, allowing comprise molecules of widely varying molecular weight, a concentration gradient to form and orientation with re- llns certainly may be the case for odors that are passively speci to that gradient to occur. Nevertheless, for most mac- released into the environment (those from plants, lor exam roscopic organisms, fluid movement has a majoreffect upon pie. may be a mixture of compounds of low and high the dispersion of chemical signals. II we imagine a simple molecular weights). However, odor mixtures (hat are ac- situation in which a fluid is (lowing smoothly over a source tively produced and emitted (such as the female-produced ol chemical molecules, the fluid is actually static at the point pheromonal mixtures of different moth species) are not of contact with the substrate. The diffusive sublayer is the highly divergent, probably because the hiosynthetic path- region closest to the substrate, where fluid motion comes to ways responsible for manufacturing the molecules have rest. This sublayer is one of several that together constitute been well conserved over evolutionary lime. what is known as the boundary layer (Denny. 1493; Vogel. Another attribute of turbulent Mow conditions is that the 1994: Weissburg. 1997. 2000}. Across the boundary layer mean direction and speed of fluid flow may be relatively there is a velocity gradient until the fluid reaches 99fr ofthe constant, bin the instantaneous (low pattern at any point in mean flow velocity (Weissburg. 1997. 2000). Boundary the plume might be difficult to predict local flow fields layers have important implications that affect both the dis- may even he heading in the direction opposite to mean flow. tribution and detection of chemical signals, and they arc. The physical conditions that result in the turbulent dispersal therefore, ofparticularimportance to animals that use chem- ol molecules therefore have several important effects that ical cues to oneiil. I or molecules that pass through the impinge upon the use of chemical signals bv animals for boundary layer, molecular diffusion continues but. depend- navigation: ing upon the conditions (such as flow speed ofthe fluid), the movement of the fluid now plays the major role in distrib- I Particles are distributed intermittently. Small-scale ed- uting molecules in time and space. In smooth, steady-flow dies can result in line-scale intermittency. whereas environments, the interaction between molecules and flow- larger scale intermittency is caused by shifts m the forms a plume that has relatively distinct boundaries, with a direction of flow and results in plume meandering. strong gradient across the transverse axis and a weaker The odor patches themselves may contain information longitudinal gradient. that can be used to orient and locale the odor source. Laminar How conditions, like molecular diffusion, are 2. Hind flow properties and their effects upon the distri- relatively uncommon .Anyone who has been fascinated In bution of odorous molecules in time and space are the patterns ofsmoke or steam issuing from a chimney and potential sources of information for navigation. being transported downwind will immediately appreciate 3. Once dispatched from a source, odors comprising ANIMAL NAVIGATION IN ODOR PLUMES 205 more than one molecule are present as a blend and refer to the ability to direct motion relative to the source of retain their ratios relative to one another, even though stimulation, e.g.. up or down a chemical gradient; kineses there may be disparities in the molecular weights of are generally reserved forsituations in which some measure the constituent molecules. oflocomotory output increases or decreases, but the resulting motion is random with respect to the stimulus source To produce changes in the behavior of an organism, (see Bell and Tobin, 1982, for review ofthese terms). This odorants must be detected by some type of chemosensory latter discussion serves to emphasize that understanding structure. Physical constraints similarto those that affect the both the locomotory activity ofan organism andthe strategy release of molecules into a fluid are equally important in for sampling the odor environment are important in accu- their detection. Most significantly, because detectors are rately describing the behavioral mechanisms utilized. housed in solid structures around which the moving fluid is In turbulent odor plumes the fluid dynamic conditions displaced, the detectorhas a boundary layer associated with play a major role in creating the spatiotemporal distribution it that odorous molecules must traverse before their capture of odor molecules. Several studies have measured odor by the chemosensory apparatus (Vogel, 1983; Loudon el<//., plume structure either in water or air to determine odor 1994; Koehl, 1996). Boundary layers around olfactory ap- parameters that have the potential to be extracted and used pendages can be affected by the speed with which the by animals that navigate in odor plumes (Murlis, 1986, organism is moving relative to the fluid orby movements of 1997; Murlis et al.. 1992; Moore and Atema, 1991: Zim- the appendages themselves through the fluid (e.g., antennule mer-Faust et al.. 1995; Finelli et al.. 1999). In a laboratory flicking in lobsters. Moore elal., 1991a). Thus, the external flume, dopamine-sensitive carbon fiber microelectrodes morphology of a chemosensory appendage and its position have been used to measure fine-scale odor plume structure relative to the animal's body (which will experience a (Moore etal.. 1989; Moore and Atema, 1991 ). Otherstudies different boundary layer) may play an important role in have utilized the same electrochemical techniques in com- determining the efficiency ofthe olfactory structure in cap- bination with fluorometric measurements (the fluorescein turing chemical cues. dye used also provides a useful visible marker) to charac- terize odor plume structure under known hydrodynamic Information Available in Odor Plumes conditions in a simple aquatic environment such as an From the brief discussion above it should be clear that estuarine tidal creek (Zimmer-Faust et al., 1995; Finelli et several sources ofinformation about the position ofan odor nl.. 1999). Measurements at fast temporal time scales show- source origin relative to the organism in question can po- that the odor plume is highly intermittent, with bursts of taesnstuiamlelsy btehatexatrrraicvtaeldaftroormmthoeveomdoerntplauwmea.yOffrocmoutrhsee,otdhoirs hceingthraotdioonr.cIonncaedndtirtaitoino,ntihnetepresapkesrsaenddwiotnhseltoswloopreszeorfo ocdono-r source is the behavioral objective of the organism in ques- transients vary systematically along both transverse and tlioocna.tiTownoofpoatnentoidaolrssoouurrccees oafreindfiosrcmuassteido:n that pertain tothe lZoinmgmietrud-iFnaaulsatxeetsaolf.,od1o99r5p:lFuimneelsli(Metooalr.e, a1n99d9)A.teTmhae,l1at9e9r1a:l edges ofthese aquatic plumes were represented by a sharp decline in the presence of both the electrochemical tracer /. Presence or absence ofthe odor anil any gradients and the fluorescein dye over a short distance. This trans- associated therewith verse gradient was much greater than any concentration The presence or absence of odor is obviously a crucial gradient along the longitudinal axis ofthe plume (Zimmer- feature ofany odorplume, and how animals respond to odor Faust et al., 1995). As distance from the source increased, or odor loss has helped us understand the mechanisms odor plumes tended to spread horizontally and vertically, animals use to navigate. In the absence of flow or visual and mean peak height and slope of odor fluctuations de- stimuli associated with the origin of odor emission, the creased (Moore and Atema, 1991; Zimmer-Faust et <//.. chemical signal itself is the only source of information that 1995; Finelli et al.. 1999). Nevertheless, much larger fluc- can be extracted. In such cases, the ability to detect the tuations in concentration than might be predicted from chemical gradient can be used to locomote toward or away time-averaged plume models were detected fardownstream from the odor source. Mechanisms known to be used for (Moore and Atema. 1991; Zimmer-Faust et al., 1995). Or- orienting with respect tochemical gradients include chemo- ganisms that orient in such plumes may use intermittency or taxis orchemokinesis andrely uponthe ability to sample the cues derived from the shape of odor pulses to steer in the milieu either spatially (comparing inputs to more than one direction of the odor source (see below; Weissburg. 1997). chemosensory structure at the same time and responding to The fluid dynamic properties ofair and terrestrial recording balance inputs more correctly known as tropotaxis) or sites result in greaterodor-plume intermittency comparedto temporally (sampling at one location, moving, then sam- aquatic sites such as an estuarine creek. Indeed, airborne pling again more correctly known as klinotaxis). Taxes plumes created by an ion source and measured in the field 206 N. J. VICKERS are highly intermittent (Murlis. 1986, 1997; Murlis el ai. several categories that can be generally applied to all or- 1992). so much sothateven in the plume the signal has been ganisms, no matter what their size and scale: estimated to be absent for 60<7r to 907c ofthe time (Murlis. 1997). Transduction 2. Odorpresence or absence coupled withfluidflow The first limitation concerns the speed with which chem- information (mechanical or visual) ical signals can be transduced by a chemosensory system. The ability to quickly process information arriving at the The presence of flow (either smooth or turbulent) pro- periphery is essential because it affects the entire response vides an additional source of information that can be inte- of the organism. For example, how frequently can the grated with the detection of chemical signals. Because the milieu be sampled and how well can sequentially arriving direction of mean fluid motion is downstream of an odor pulses beresolved (i.e.. howclosecan twopulsesbe without source, any odor that is detected must have come from being perceived as a single pulse)? These sampling param- upstream. By resolving direction of fluid movement, an eters may be particularly important for most macroscopic organism can determine what direction to take towards a organisms whose olfactory environment includes turbulent distant and frequently unseen source of odor. The mecha- plumes and intermittency as normal features. The threshold nism for the detection of flow depends, to a large degree, ofresponse may also be an important constraint in respond- upon the life history of the organism in question. For ing to chemical signals. A certain number of molecules of animals that dwell on a fixed substrate, flow can be detected the same type may have to arrive simultaneously (or within by mechanical means i.e., pressure differences detected the integration period ofthe sensor) toproducearesponse in by mechanoreceptors located on the animal's body, head, or the sensory apparatus. Even the simplest systems are limited other appendages. The direction of flow can be gauged by by the speed with which odorant molecules are cleared from the deflection of mechanosensors with reference to the their receptors (Denny, 1993). Molecules that occupy a substrate. Detection of flow in this manner can be compro- binding site effectively prevent a receptor from responding mised by locomotory activity that will also stimulate mech- to future encounters with odor. Thus rates ofsensory adap- anoreceptors. For this type of flow detection to operate tation and disadaptation certainly influence the ability ofan efficiently, the organism must be able to distinguish move- organism to respond to repeated chemical stimuli. In ani- ment of the flow due to the flow itself from activity in mals with even simple nervous systems these peripheral mechanoreceptive cells resulting from locomotion. In other processes may have higher order corollaries such as habit- substrate-bound organisms, flow direction can be estimated uation and learning that can affect the behavioral perfor- by differential cooling: a dog's nose and your own finger, mance. wetted and held in the air, are instruments for detecting wind direction. Animals that orient to odor sources while Processing flying or swimming are faced with a somewhat more complex task. Even though mechanical detectors can be stimu- The second important biological limitation is how lated by the activities of flying or swimming, the need to quickly information is processed by the organism. Process- discriminate between movement ofthe fluid due to flow and ing may or may not involve communication with othercells. movement of the fluid due to self-propulsion through it For example, transduction of chemical signals by a unicel- means that mechanoreception may not be a particularly lular bacterium in a chemical gradient may not be followed reliable source of information for determining flow direc- immediately by a response, because otherbiochemical path- tion. Vision is far more reliable because locomotion can be ways have to be activated before the cell changes direction referenced to fixed points in the environment such ax the or rate of locomotion. Similarly, in organisms with a mul ground pattern or riverbed. In fact, experiments with flying ticellular nervous system, the existence ofchemical connec- moths and other insects have shown the extraordinary de- tions between neurons imposes an mesilaMc dcla\ between gree to which these creatures rely upon visual feedback in transduction events at the periphery and the resulting be- making theirway upwind in an odorplume (Kennedy. 1939: havioral response. Thisdelay affects the rapidity with which David. 1986; Colvin el til.. 1989). an organism can respond to the chemical environment. I'.ioln'jK .il Limitations Behavior In addition to the constraints imposed by the physics of Behavioral limitations include constraints imposed by the the fluid medium in which an organism exists, many bio- organism itself. There are obvious gross restrictions which logical constraints must be included in any understanding of include, forexample, only being able to walk or run instead how animals navigate and orient with respect to odor of fly. In addition, there are finer locomotory limitations In plumes. These biological rcsinciions can be broken into cither words, a bacterium can move only so many body ANIMAL NAVIGATION IN ODOR PLUMES 207 lengths or make directional changes only so many times A. Crab every second; each species of moth has a maximum sus- 40 tainable flight speed, etc. Ofcourse, there will be variation CQO)HJT 20 among individuals, but each organism has an upper limit. II u- These behavioral effects are not independent of delays tn ra causedby the transduction andprocessingevents mentioned O2-O -20 -40 above but are inclusive of them; together they can be considered as a behavioral latency. Taken together, these B. Lobster biological constraints influence the speed with which organ- 40 isms perceive and react to changes in the chemical environment. Our own ability to notice such changes is constrained by our ability to resolve differences in the behavioral output -20- oforganisms. Marked changes can be readily observed, but '- more subtle variations in behavior may reflect important 50 100 150 200 250 300 fine-tuning oflocomotory activity that might easily be over- Downstreamdistance(cm) looked with observational techniques that do not sample the C. Moth behavior at a sufficiently fast frequency. 20 10 Mechanisms of Navigation in Odor Plumes 0- -10- One cannot help but notice that dissimilarorganisms tend -20 to have similar paths in solving odor-orientation problems 50 100 150 (Fig. 1). Crabs and lobsters pursue somewhat meandering Downwinddistance(cm) upstream paths when stimulated by appropriate odors (Fig. Figure 1. Different organisms frequently have similar paths through 1A, B). A male moth orienting to a calling female or a space during odor-mediated orientation. However, the similarity in trajec- synthetic sex-pheromone source (Fig. 1C), a procellariiform tories does not necessarily mean that each animal is employing the same obirridenhtionmgintgo iintsonnataivpeotesnttriealamsoaullrcemuosftfomoodv,eanadgaainssatlmtohne aoarnnideondttoharetiscoornuabrm'cesechcparoesnaiittsiemod.nbyi(sAa)mcaAlrakcmerdaibn,ataCta1il-dlsailiiincetrceetreevksa.lssFalp(iofdwruosvm,elZooircimietmnyetriis-ngF1a0tuocswmta/rsed,t mean direction offluid flow, and all make turns during their al., 1995, withpermission). Theboundaries ofthe plume as visualized by upstream progress (the latter two examples are illustrated in fluoresceindyeareindicatedbytheshadedarea.Thecrabisthoughttouse fig. 1 from Arbas et al., 1993). This observation is based a combination ofrheotactic andchemotactic orientation mechanisms. (B) upon the similarity of the tracks that the animals follow Tenhiezeldobmstuesrs,elHoimnaarnuesxpaemreirmiecnatnauls,flourmiee.ntFilngowtovealnocoidtoyrwsaosuracpeproofxihmoamtoegl-y through space as they navigate toward an odor source, 1 cm/s.andthelobster'spositionismarkedat 1-sintervals(fromMooreet although we must bear in mind the important distinction al.. 1991b. with permission). The results of this study suggested that between those creaturesthat fly or swim freely through their lobstersusedachemotropotacticorientationmechanism.(C)Flighttrackof fluid surroundings (moth, fish, bird) and those that orient a male moth. Heliothis virescens. responding to a point-source offemale while in contact with the substrate (crab, lobster). Even for sex-pheromone.Thewindspeedwas60cm/s.Notethatthepositionofthe animals in similar situations, we cannot assume that the mMoatlheimsomtahrskehdaveeverbyeeln/3s0thhows,nastocoumspearaedvitsouatlhleycgruaibdaenddmleobcshtaernitsramckst.o behavioral mechanisms that underlie the production oftheir compensate for wind-induced drift and steer upwind (optomotor anemo- tracks are the same, even though this assumption is a taxis).Thereversalsfromone side ofthe wind line to the otherthatoccur tempting one. Instead each organism must be viewed within along the track are the result ofa second program ofinternally generated the context of the particular conditions under which the tmuercnhsankinsomwsnaraesmocdouulntaetretdurbnyincgo.ntaBcotthwitahneodmoort.acItnicallatnhrdeecsoiutnutaetriotnusr,nitnhge behavior occurs naturally, because these are likely to rep- directionoffluidflowisfromlefttoright(asindicatedinA).Notethatthe resent the pressures that have shaped the behavioral and scales for A and B are the same, whereas C is enlarged. In A and B, the transduction/processing mechanisms over evolutionary odorsource waslocatedatthe graph origin (X = 0, Y = 0). whereasinC time-scales. thepheromone sourcewaslocatedabout 30cmfurtherupwind,tothe left, Nevertheless, in each of the examples above it seems ofthe graph origin. possible that the fluid environment provides either visual or mechanical stimuli that might be integrated with odor information to provide some measure ofthe direction offlow. 1997; Willis and Arbas. 1997; Vickers, 1999). Thus an The upwind flight of male moths under the olfactory influ- extensive review of olfactory-mediated flight in moths ence of female-emitted pheromone has been particularly seems unnecessary here. Only relatively specific comments well-studied and extensively reviewed recently (Baker and will be made, while integrating the moth system with those Vickers, 1997; Carde and Mafra-Neto, 1997; Kaissling, of other animals that orient with respect to turbulent, inter- 208 N. J. VICKI ks mitteni odor plumes to point out common features of the appears to occur with the animal moving up-current. How- navigational mechanisms. ever, in only one species has experimental evidence been accumulated that indicates that up-current movement in the Integratingflow and odor detection picscnce ofodorrelies upon visual feedback from objects in the environment (Emanuel and Dodson. 1979). Sharks are As was mentioned earlier, the most reliable method for also able to locate sources of odor, perhaps by using a animals that acti\el\ fly or swim to detect direction offluid combination ofrheotactic (an oriented response with respect flow is to reference movement relative to stationary objects to the direction ofwater flow) and chemotactic mechanisms in the environment. This has long been known to occur in (Hodgson and Malhewson. 1971). insect-., as Kennedy ( 1939) first demonstrated with mosquitoes flying in a plume constituted ofhis own exhaled breath. Much subsequent work on Hying male moths orienting in a pheromone plume has shown that these animals are ex- Anothercommon feature ofthe tracks ofanimals that fly. tremely sensitive to movements of the ground pattern (for swim, or walk to odor sources is that up-current progress example, see Baker and Vickers. 1994). l-issentially. the rarely occurs in directly straight lines (Fig. 1 ). Most animals fluid movement causes a discrepancy between the animal's observed thus far lend to switch back and forth across (In- direction of movement (heading or course) in the fluid and direction of fluid flow. In flying inoihs. this phenomenon static objects in the environment. Resolution ofthe resultant has been well-studied and the mechanism underlying it drift is hypothesi/ed to enable moths to steer more or less debated (Baker and Vickers. 1997: Carde and Mafra-Neto, into the wind and also to ascertain whether they are making 1997: Willis and Arbas. 1997: Wu/gall. 1997). A numberof progress against the wind or are "losing ground." i.e., being experiments, which have been summari/cd in greater detail carried downwind (Baker and Vickers, 1997; Carde and (Baker and Vickers. 1997: Carde and Mafra-Neto. 1947: Mafra-Neto. 1997: Kaissling. 1947: Willis and Arbas. 1997: Vickers. 1999). suggest that the species-specific turning Vickers. 1999). rates are driven by an internal "clock" that results in regular The e\idence for such visually steered mechanisms in reversals from one side of the wind line to ihe other as the other animals is not yet as convincing, largely because the moth makes net progress toward the source. The mecha- data derive from observational rather than manipulative. nism, known as counterturning. is integrated wiih the visu- experimental studies. Observations of the flight trajectories ally steered compensation for wind-induced drift: together and behavior ofprocellariiform buds over the open sea has these mechanisms produce the upwind /ig/agging flight shown that they approach sources of odor produced by tracks typical of many insects flying to an odor source (Fig. slicks from downwind, presumably using their extremely 1C). Other animals such as birds (Arbas </ <//.. 1993) per- well-developed olfactory systems (Nevitt. 1999). Perhaps form similar maneuvers during upwind flight in response to some sort of visual feedback is utili/ed to determine wind odor, but again, the actual mechanisms underlying the be- direction. However, it is not clear how pelagic buds inte- havmrhave not been examined underrigorous experimental grate visual cues with olfactory signals, given that the entire conditions. Likewise, tish such as salmon were observed to ocean is heaving and swelling and otherwise certainly not /ig/ag as they returned to their home streams (Johnsen and stationary. The possibility exists that these birds exploit Hasler. 19X0). some other mechanism to determine wind direction. The Suhstraie-dwelling animals that inhabit aquatic environ- slick odor sources used in such experiments must be essen- ments have been the locus of many studies upon odor- tially invisible to an approaching bird unless other individ- mediated behavior. Fluid dynamics are somewhat easier to uals have already arrived at the site, thereby providing measure and manipulate in the aquatic realm, and so these visual cues themselves. Thus, the birds that first arrive are studies are particularly enlightening as to the relationship presumably the tines most capable of locating the odor between fluids, odors, and the potential mechanisms em- source through olfactory signals alone or through olfaction ployed by animals lo locate odor sources. Both crabs and integrated with visual/mechanical cues, and future studies lobsters have been the subjects of intensive investigation should focus on those species In other tield studies, petrels (Weissburg and Zimmer-Faust. 1993. 1994: Zimmer-Faust and shearwaters were observed to approach food sources, el nl.. 1995: Moore et /.. 199lb). For the blue crab. Ctil- nesting colonies, and individual nesting sites under cover of linectex sapidm. flow is an important cue in determining darkness and alw inti downwind (Grubb. 1972. 1974. successful orientation to an odor source. No-flow or high- 1979). Under such conditions, darkness may preclude the flow conditions resulted in poor orientation responses, use of visual feedback for the purposes of determining whereas a slow flow speed resulted in high levels of source wind-induced drift, again raising the possibility that other location (Weisshurg and Zimmer-Faust. 1993). These ex- mechanisms are employed. periments suggested that males made orientation responses In many species of tish, odor-modulated navigation also with respeci to ihe direction offlow (rheotaxis). In addition. ANIMAL NAVIGATION IN ODOR PLUMES 209 these crabs also used chemotaxis, as evidenced by their should reveal to what extent odor plume dynamics are turning behavior upon crossing the relatively sharp lateral altered by locomotory activity or movement of sensory edges of the plume. The turning behavior of C. supitlus is appendages, an important consideration given that most of therefore distinct from that of flying moths, which respond the measurements of plume structure to date have used to an internal turning clock rather than to the detection ofa stationary detectors. lateral chemical gradient. The lobster Honi(ini\ ainericanits also responded to odor plumes but appeared to rely more Orientating in "restricted" sensory environments upon the information contained within and between odor pulses for orientation than upon the direction of flow itself Experiments that alterthe sensory world ofan animal can (Moore et al.. 1991b). Thus, these animals may use simul- be informative as to the mechanisms that those animals use. taneous bilateral comparisons to remain or turn into the For example, moths forced to fly through a visually depau- plume and sequential comparisons of odor intensity to ad- perate "blank" tube were unable to remain oriented with vance toward the source (Weissburg. 1997). respect to a pheromone plume compared to situations in which the visual environment was enriched (Vickers and Recording odorplume dynamics and behavior Baker. 1994b). Baker and Keunen (1982) stopped the wind for flying oriental fruit moths, leaving a stationary plume, simultaneously suspended in the air. Deprived ofthe wind-induceddriftthat In general, the study offluid dynamics in air is somewhat is thought to be necessary for steering upwind, some males less tractable than in water. One of the ways that our were still able to locate the odor source, suggesting that knowledge of orientation mechanisms and odor plume dy- other mechanisms can be invoked under certain conditions. namics has been advanced in recent years is by utilizing Some animals, however, exist in environments that afford techniques that can detect odors or odor representatives. only limited sensory feedback. Such situations might pose Murlis and co-workers used ion generators and detectors to navigational problems for these creatures. How, for exam- measure odor plume dynamics in terrestrial habitats (Mur- ple, might fish or whales orient with respect to chemical lis, 1986, 1997; Murlis et al.. 1992). These studies have signals in mid-ocean where stationary visual reference been particularly instructive in establishing the types of points are unavailable because everything drifts in the same odor stimuli that might be encountered by flying insects, for current? How do deep-seadwellers orient where visual cues example. are unavailable due to the lack of illumination? How do The actual olfactory appendages ofinsects have alsobeen pelagic sea birds reference wind direction with respect to successfully used to determine odor plume structure and the rolling ocean, moving beneath them? What do insects in dynamics. The moth antenna is easily detached and can be dark forests or cave-dwelling animals do when chemical readily hooked up to electrodes. Electrical activity mea- signals beckon, but visual references are unavailable to tell sured across the antenna is thought to represent the com- them the direction of flow? Presumably, chemical signals pound activity of the olfactory receptor cells housed in still have arole to play in all these situations. Atema (1995) cuticularprojections that often coverthe antenna) surface by has suggested that some animals might be able to use "eddy the thousand. The resulting electroantennogram (EAG) has chemotaxis" to orient with respect to chemical cues. In this been useful in visualizing pheromone plume structure. case, the actual mechanical deformations in the fluid that are Baker and Haynes (1989) used EAGs to measure plume associated with the chemical signal are important in guiding structure in the field and pushed an EAG preparation along locomotory activity. It is certain that mechanical deforma- a wind tunnel toward a pheromone source to simulate the tion and vortices associated with otherorganisms can play a odor plume that a flying male encounters (Baker and Vick- significant role in altering the behavior ofother organisms. ers, 1994). An EAG preparation transported by an intact Recently. Yen and co-workers (Doall et al., 1998; Weiss- flying male gave an impression of the frequency and mag- burg et al.. 1998; Yen et al.. 1998) demonstrated that male nitude ofolfactory stimulation that a male moth encounters copepods have the remarkable ability to quite literally track as he flies upwind in a pheromone plume emitted by a down conspecific females by following their wake struc- single, point-source (Vickers and Baker, 1994a). Lobsters tures and possibly the chemical signals contained therein. transporting an electrochemical microelectrode and "back These tiny planktonic creatures also engaged in "casting" pack" gave a similar impression oftheir olfactory world at behavior when the mechanical/chemical trail went cold. the olfactory receptor level (Basil and Atema, 1994). These There are other senses too, that may conceivably be inte- studies are important because they have allowed us to grated with chemoreceptive input to provide a sense of visualize not only the dynamics of the odor plume but also direction toward an odor source. For example, electric fish the kinds of information that are being detected by the are able to "see" theirenvironment: maybe they can use this actual peripheral olfactory sensors and transmitted to the sense to resolve deformations of the fluid world that they brain for further processing. Moreover, studies of this type inhabit and couple this information with chemical input to 210 N. J. V1CKHRS orient toward an odor source. The lateral line of fish may number ofodor pulses available (Vickers and Baker, 1996; provide mechanical cues associated with fluid conditions. Vickers. 1999). Thus both mechanisms are directly modu- Perhaps animals can "visualize" fluid flow by means of lated by odor encounters. These findings have been re- sonarorthermal sensory systems, particularly in the aquatic viewed extensively recently (Baker and Vickers. 1997; realm. In short, any sensory modality that can provide an Carde and Mafra-Neto, 1997; Vickers, 1999). indication of some attribute of the fluid environment is Relatively little is known about how other organisms likely to have been exploited in one way or another. At respond to odor gaps, or even about how long a time present our search image is somewhat biased by those between odor stimuli is necessary to trigger the adoption of sensory modalities that we are most attuned to. namely a new strategy. A simple strategy to handle gaps in odor vision and sound, but animals never cease to amaze us h\ stimulation is to arrest upstream movement and wait for either the extraordinary heights to which some of their olfactory stimulation to resume. This would seem to be senses are adapted orthe existence of"sixth" senses that we particularly advantageous when the flow is relatively steady simply do not possess. (producing plumes that do not meander too much), the organism is a substrate dweller, and the target emitting the Changing strategies as a result ofmoving in and out of signal is relatively sessile. Thus, a blue crab tracking the odor or how to span odor gaps'.' p(Zliummmeere-mFiattuesdt ebtyala.,s1e9d9e5n)t.arMyotmhusssdeol niostaswgiotocdh teoxaamnpelwe Intermittency is a feature of any odor plume that is mechanism but adjust both anemotactic and counterturning generated in a turbulent fluid environment (Murlis. 1986, programs to result in tracks that move side-to-side perpen- 1997; Murlis el al, 1992; Moore and Atema, 1991; Zim- dicular to the direction ofmean flow. This behavior, known mer-Faust et al.. 1995; Finelli et al., 1999). By moving in as casting, is viewed as a way for the animal to resume andoutofturbulent plumes, organisms are likely to increase contact with the odor plume. Casting in H. virescens in- intermittency. producing even greater fluctuations in the volves increased vertical as well as lateral excursions, chemical signal. From an animal's perspective, an important thereby facilitating a three-dimensional search of space issue revolves around the decision ofwhen to alter behavior (Vickersand Baker, 1996). Undercertaincircumstances, the in response to odor onset and loss (or perceived loss). animals will eventually drift back downwind; videorecord- Failure to act appropriately may reduce the chances of a ings of oriental fruit moth tracks in the field have shown timely arrival at a source ofodor. On the other hand, some that, on occasion, these insects will actually pursue a down- gaps can probably be ignored because they do not represent wind course, perhaps in a effort to regain the odor plume relevant changes in the stimulus and responding to them (Baker and Haynes, 1997). Once contact is re-established, would reduce search efficiency. Thus animals must be able the insect may resume upwind flight, trying again to locate to determine how quickly to respond to gaps in odor and the source. what kinds of changes in stimulus parameters (e.g., frequency orconcentration) or shifts in flow environment must Future Directions eolfictihteapruelsspatoinlseesatnrductwuhraetokfinoddsorcapnlbuemeisgnuorpeodn. mTohtehefffleicgthst relCalteiaornlsyh,ipthbeerteweaerne frleuliadtidvyelnyamfiecsw,soydsotrempsluimnesw,hiacnhd atnh-e behavior have been investigated. Vickers and Baker (1992) imal orientation have even begun to be worked out. Onecan determined that male H. virescens required the generation of only hope that future endeavors will include organisms of a certain number of pulses per second in order to sustain varying scale that inhabit all types offluid environment. On upwind flight. At a production frequency offewerthan four the basis of the various mechanisms animals employ to filaments per second, males were unable to locate the odor locate odor sources, I suggest that the following areas source. Subsequent studies revealed that males are actually should prove fruitful for future research. capable of responding to encounters with individual filaments ofodor and, after intercepting an odor filament, they The olfactory environment make brief upwind surges towards the source (Vickers and Baker. 1994a, 1996). Other experiments using males of a Much of our understanding has come from studies in distantly related moth species. Cadm cautella, revealed that which discrete odor sources have been used as the source of this species also responds to encounters with single strands stimulation in relatively simple fluid environments (labora- ofpheromone (Mafra-Neto and Carde. 1994. 1996). In both tory flumes and wind tunnels, tidal creeks and open fields). species, the frequency of filament generation has a signifi- The actual environmentsorganisms inhabit are considerably cant effect upon the directness of upwind flight: the faster more complex in olfactory content and fluid dynamic prop- the frequency, the straighter the flight. Measurements on // erties. For example, the trajectories and structure of phero- virescens revealed that both the anemotactic component ol mone plumes created in a forest are different from those in upwind flight and the counterturning were affected by the an open field (Elkinton ct al.. 1987; Willis ft al.. 1994). ANIMAL NAVIGATION IN ODOR PLUMES 211 How are an animal's orientation responses modulated in ened the notion that animals are efficient at locating sources morecomplex fluid mechanical situations? Are novel mech- of odor in their habitats. Their orientation responses are anisms employed? Understanding how the cues available in appropriate for the types of odor plume that they typically such landscapes are sorted, discriminated, and processed encounter. Ourcurrent understandingofalimitednumberof will provide insight into the behavioral mechanisms that model systems in relatively simple settings provides a solid animals use in their habitat. foundation for future endeavors to understand the vital role that chemical signals play in mediating many of life's Behavioral capabilities in natural arenas interactions. We still know very little about the behavioral capabilities ofanimals. How faraway from an odorsource, forexample, Acknowledgments will a male moth orient, and how do the behavioral mech- Thanks are due to Dick Zimmer for the invitation to anisms change as the male approaches the source? Recent participate in the symposium at the Western Society of field experiments have shown that males generate faster Naturalists meeting in December 1998 and for his great airspeeds and counterturn less frequently farther away from enthusiasm and positive criticism of an earlier version of a pheromone source (9-11 m) (Vickers and Baker, 1997). this manuscript. For practical reasons, laboratory studies in wind tunnels or flumes examine the behaviorat 1-2 m fromthe source. How Literature Cited do the fluid dynamics affect the distribution and shape of odor pulses at greater spatial and temporal scales? What are Arbagsoa,l-Eo.riAe.n,teMd.loAc.omWoitliloins,: apnhderRo.moKnaen-zmaokdiu.la1t9e9d3.fligOhrtgabneihzaavtiioorn ooff the spatial, temporal, and fluid dynamical limits of behav- moths. Pp. 159-198 in Biological Neural Nem-orks in Invertebrate ioral responsiveness? These questions remain largely unan- Neuroetho/ogv and Robotics, R. D. Beer, R. E. Ritzmann. and T. swered. McKenna, eds. Academic Press. San Diego, CA. Atema, J. 1995. Eddy chemotaxis and odor landscapes: exploration of nature with animal sensors. Biol. Bull. 191: 129-138. Odorplume dynamics Baker, T. C., and K. F. Haynes. 1989. Field and laboratory electroan- Locomotion. It is clear that the locomotor tactics of ani- tennographic measurements ofpheromone plume structure correlated mals can change the perceived structure of an odor plume. Bakewri,thT.orCi.en,taalndfruKi.tFm.otHhaybneehsa.vi1o9r.96.PhysPioh/e.roEmnotnoem-olm.ed1i4a:te1d-1o2p.tomotor Stationary measures of odor plumes have been and will anemotaxis and altitude control exhibited by male oriental fruit moths continue to be instructive. However, an animal must move in the field. Physiol. Entomol. 21: 20-32. in order to locate an odor source. To what extent are Baker,T.C.,andL.P.S. Kuenen. 1982. Pheromonesourcelocationby stationary measurements of odor plume structure affected flying moths: a supplementary non-anemotactic mechanism. Science 216: 424-427. bsyamientkriondduscionfgomdoovrepmuelnste?flDucotutatwioonsmoitfhosneexipser1iemnceawtahye Bakemra,leT.moCt.h,satondphNe.roJm.onVeicfkielrasm.en1ts99a4n.dvBiseuhaalvisotrimaullir.ePapc.ti8on38t-i8m4es1oinf m fromanodorsource and the otheris20 away, flying twice Olfaction and Taste XI. K. Kurihara, N. Suzuki, and H. Ogawa. eds. as fast, and turning less often? Springer-Verlag, Tokyo, Japan detNeecutreodbiaonldogrye.prWesheanttefdeabtyurpeesriopfheordaolrapnldumceentsrtarlucntuerrevoaurse BakeCmroa,trhdsTe..aCPn.pd,.Aa2.n4d8K.-N2.M6i4Jn.kisnV,iPcekhdeesr.rso.Cmho1an9p9em7.aRenseaPanhrdecrhHo:amlolN,neeNw-emDewidriYeacottriekod.nsf,ligRh.tTi.n systems? Some studies have used neurobiological prepara- Baker, T. C., B. S. Hansson, C. Lofstedt, and J. Lofqvist. 1988. tions such as EAGs (Baker and Haynes, 1989; Vickers and Adaptation ofantennal neurons is associated with cessation ofphero- Baker, 1994a) or receptor neuron recordings (Baker et al, mone-mediated upwind flight. Proc. Null. AcaJ. Sci. U.S.A. 85:9826- 1988; Van der Pers and Minks, 1993) to track peripheral 9830. Basil, J., and J. Atema. 1994. Lobster orientation in turbulent odor events associated directly with actual, behaviorally relevant plumes: simultaneous measurement oftracking behaviorand temporal odor fluctuations instead of with odor tracers or ions. Fur- odor patterns. Bio/. Bull. 187: 272-273. ther experiments that combine neurobiological measure- Belanger, J. H., and M. A. Willis. 1996. Adaptive control of odor- ments of actual odor with detection of electrochemical or guidedlocomotion:behavioral flexibilityasanantidotetoenvironmen- ion tracers (e.g., Murlis el al., 1990) could be extremely Bell,talWu.npJr.e,diacntdabiTl.itRy.. ATdoabpitn..B1e9h8a2v.. 4C:h2em1o7--o2r5i3e.ntation. Bio/. Rev. 57: informative in determining those features of filamentous 219-260. odor plumes that are extracted by the nervous system. Colvin, J., J. Brady, and G. Gibson. 1989. Visually-guided, upwind These suggestions are, ofcourse, not exhaustive, but they turning behaviour of free-flying tsetse flies in odour-laden wind: a do seem attainable in the foreseeable future. In conclusion, wind-tunnel study. Physiol. Entomol. 14: 31-39. tthiemeimapnodrtsapnaccee ohfasflculiedardlyynabmeiecnsdeinmodnissttrraibtuetdinignotdhoersfeiwn CardDmieor,tehcRst.ioTtn.os,,paRhn.edrTo.Am.CoanMread.feraaP-pn.NdeAt2.o7.5K.-129M99i07n.ksin,MePedhsc.ehraConhmiaospnmmesanoRfeasfnlediagrhHtcahlol:f.mNNaeelwwe systemsthat have been thoroughly studied. A growing num- York. berofobservational and experimental reports have strength- David,C. T. 1986. Mechanismsofdirectional flightin wind. Pp.49-57 212 N. J. VICKF.RS in Mechanisms in Insect Olfaction, T. L. Payne. C. E. J. Kennedy, and Murlis. J.. J. S. I Ikiiiinii. and R. I. Carde. 1992. Odor plumes and M.C. Birch,eds. Clarendon Press.Oxford. I'inled Kingdom. how insects use them. Anna. Rev. Entomol. 37: 505-532. Denny. M. W. 1993. Airand Water: the BiologyandPhysics ofLife's Vvitt, (i. 1999. Foraging by seabirds on an olfuclory landscape. Am. Media. Princeton University Press. Princeton. NJ. V , S7: )(, 53 Doall. M. H.. S. P. Colin. J. R. Slrickler. and J. Yen. 1998. Locating \all der PITS.J. N. ('..and A. K. Minks. 1993. Pheromonemonitoring a male in 3D: the case of Temora long/comix. Philos. Trans. R. Soc. in the lick! using single scnsillum iccoidmg Entomol. Exp. Appl 68: Land. B353: 681-689. 237-245. Klkinlnn. .). S.. C. Schal. I. Onn. and R. T. Carde. 1987. Pheromone \ickcrs. N. J. 1999. Ilu- ellects of chemical and physical features of putt trajectory and upwind (light of male gypsy moths in a lores! pheiomone plumesupon thebehavioral responsesofmoths. Pp. (V> Its Physiol. Entomol 12: 106 in \J\tiiitt . in Chemical Sii;nal\ in \'ertehrale.\. R. F. Johnston. D. K.manucl. M. K...and.1.J.Dodson. 1979. Modificationoftherheotropic Mullei-Schvvar/e, andP. W. Sorenson.eds. KluwerAcademic/Plenum behaviorofthe male rainbow trout(Salinogairdneri)by ovarian fluid Publishers. New York. J. Fish. Res. BoardCan. 36: 63-68. \ickers. N.,)..and T. C. Baker. 1992. Male lleliotltisvirescenssustain Finelli,C.M.,N.D. PentchefT. R. K./.immer-Faust.and 1).S. Welhey. upwind thghl in response lo experimentally pulsed filaments of their 1949. Odor transport in turbulent Hows: constraints on animal navi- sex-pheromone. ./. ln\fct Behav. 5: 669-687. gation. Limnol. Oceanogr. 44: 1056-1071. \ickers. N.J..and T. C. Baker. 1994a. Reiterativeresponsestosingle druhh. T. C. 1972. Smell and foraging in she.iiwaters and petrels. strands of odor promote sustained upwind Hight and odor source Nature 237: 404-405. location by moths. /';,. \,,il. !,,/. Sci. U.S.A. 91: 5756-5760. (.ruhl), T. C. 1974. Olfactory navigation lo ihe nesting burrow in \ickers. \. ,|.. and T. C. Baker. 19941). Visual feedback in thecontrol Leach'spetrel(Oceanodromaleucoiilu>a>..\niin. Bchav. 22: l'2 202 ofpheiomone mediated (light ol ll<liatln\ UM-WI/H males (I.epidop- (iruhh. I. C'. 1979. Olfactory guidance of Leach's si,urn petrel to the tera: Noctuidael. J. ln\i\t lichav. 7: 605-632. JHoohdnbgisnrseeosnenhd.a.irnPkH.gs..BiS.A.s..lnaanan.ndn.ddNWAYRi..lA1sK)co..anMHdauBsutlSlhcelifr..u9s111n8:98n8:.10.411179-471T-14h.13e82u.Csheeomfocsheenmsiocraylocruieesntiantitohne \'icAt(kloeirg1ish7n,t8t:eNHr.ac8ce.3)pk..t1ia-so8hnna4dp7oefI.ifnCi.lHaemBleainiktiestrh.iosf1n9s9Ve(cx>..Mphi'e1\raotiemlnocniieemsalaoenfsd.becJhl.aevaCnioomraiparl.irPnehfsylpsuoienonslc.ee upstream migrationofcohoslamon. Oncorhyncus knuitIt\\'albaum../. Nickers. N. J.. and T. C. Baker. 1997. Flight ol Hclunhis virescens Fish Biol. 17: 67-73. males m ihe held m response lo se\ pheromone. Physiol. Entomol. 22: Kaissling. K.-K. 1997. I'heromone-controlled ancmotaxis in moths. 1'p. 277-285. 343-374inOrientationandCommunication inArthropods. M. I.ehrer. Vogcl. S. 1983. How much an passes through a silk moth's antenna' ed. Birkhauser V'erlag. Basel. Svvii/erland. J. Insecl /'/MX/,-/ 29: 597 1,112 Ken/n.e>d!y,SoJt. S.Lo1n9d3o9.nAT1h0e9:us2u2a1l-2r4e2sponses of living mosquitoes. /'/<>, Vofiel, S. 1994. /.//C in Moving IInuh. 2nd ed. Princeton University Press. Princeton. NJ. Koehl. M. A. R. 1996. Small scale fluiddynamicsol olfactoryantennae. XVeisshurg. M. J. 1997. Chemo- and mechanosensory orientation by l.ouMdaorn..FC.r.esBl.mA.ItB,e-lsi,ti\,.a/n'/dMS,M;../.A.27:R.1K2o7eh1l4.1.1994. When does motion crustaceans in laminar and turbulent Hows: liom odor trails to vortex Mafrr/ae-l/aNtiri'vteol.tion.Anl.e,i1ga9hn3b:dor2Ri.Wng2'I5s.4mC.laarcdees..il1i9er94t.he fFlionwe-ishcraloeugshtraurcrtauyrseooflhpaihrcsi'.1 WeissMts.rneeultr.seu.h.iIe'ipM..2ed1J.4. B22i0r40k06h.ainusOTerhrieeVnetfrlalut.iiidgo.ndyaBnanasdemliC.coamSlmwucinoiin/etcrealxtatnidoo.nfienhAermtohsreonpsoodrsy. omone plumes modulates upwind orientation of Hying moths. \'aiitre behavior Hiol Hull 198: 1S8-202. 369: 142-144 \Veisshiirg. M. .1.. and R. K. /.immer-Faust. 1993. Life and death in \lafra-Ncto, A., and R. I. Carde. 1996. Dissection ol Ihe pheromone- moving fluids: hydrodynamic ellecis on chemosensoiv mediated pre- modulated Might of moths using single-pulse response as a template dation Ecology 74: 142S 1443 Experientia 52: 373-379. \\eisshurn, M. .1.. and R. K. /.innner-Fanst. 1994. Odor plumes and Mimrsieo,naPl.fAi.n,eastnrduc.t1u.rAetotfmaa.n a1q9u9a1t.ic oSdpoatriaplluimnteorHmiantti.onHuilnlth1e81t:hreIeO-Xd-i4mIe8n.- Weishshoiwirjb-l.ueMc.rab.s]..usMe.thIeI.m Dionafliln,diangndpreJy..YJ.enE.\p.19H9i8o.l 1f9o7l:lo3w4i9n-g37t5h.e MiKirv, P. A., (i. A. (ierhardt, and .1. AU>ma. 1989. High resolution invisible trail kinematic analysis of mate-tracking in the copepod spatio-temporal analysis ofuqiuik ihcmical signals using micioelei li-nniiii luimitornis. Philos. Trans. R. Soc. Limtl. 353: 701-712. trochcmical electrodes ( Inm Senses 14: X2') X40 \\illis. M. A., and K. A. Arnas. 1997. Active behavior and reflexive Moon-. P. A.. .). Aterna. and (i. A. (ierhardt. 1991a. Fluid dynamics responses: anolhei peispective on odor-modiilaled locomotion. Pp. and microscale chemical movement in the chemosensoiy appendages ill! M1' in rhi-nnnunc Research: New Directions. R. T. Carde and ofthe lobster Hinnnriis umi'ritttnns. Chem. .SV/n< s Hi: (td3-674. A. K. Minks, cils. Chapman and Hull. New York. Moore, P. A.. N. Schol/.and .). Alema. 19911). Chemical orientationof Willis. M. A.. ( . I. David, ,|. Murlis, and R. T. Carde. 1994. 1Heels lobsters. Homarus ameinuniis. in turbulent odor plumes. J. Cln-m ol pheromone plume structure and visual stimuli on Ihe pheromone- Ecol. 17: 1293-1307. moilul.iied upwind flight ofmale gypsy moths (Lymantriadispan in a Murlis. J. I98<>. The structure ..I odnui plumes I'p 27 (S m \l,, /,,, forest (l.epidoplera: l.ym.iniindaci ./. Inscit. Hchav. 1: 385-40'). nisms in Insect Oll,n H,H\ I I I'.iyne. C'. F. J. Keimedv .mil \1 ( Wit/nail. P. 1997. Modulation ol pheiomone-medialed flight in male Birch, eds. Clarendon Press. Oxford. I niicd Kingdom moths. Pp. 265-274 in I'/unnni'in Research: New Directions, R. T. Murlis.J. 1997. Odorplumesandthesignalstheyprovide. Pp. 2'I >31 < .ml. .mil A K. Minks, eds. Chapman and Hull. New York. inPhenimtmeResearch:NewDirections. R.T.CardeandA K Minks. Yen..1.. M..1. Weisshiirg,and M. II. Doall. 1998. The fluid physicsof eds. Chapman and Hall. New York signal perception by mate-tracking copepods. Philos. Trans. R. Soc. Murlis..|..M.A.Willis,and K.T.Carde. 1990. Odoursignals: patterns Loud 353: 7S7 SIM, in time and space. Pp. 6-17 in Proceedingsofthe Tenth International /.immer-l anst.R. K.,C.M.Kinclli,N.I).PenkhelT.and I).S. Wcthey. Symposium on (lll,nimn mid In-.li K H Moving, ed. draplui I "in 1995. odor plumes and animal navigation in turbulent wuter flow: a municalion Systems. Oslo. \mwav field study. Biol. Hull 188: 1 I I 116.