272 REPORTS FROM THE MBL GENERAL SCIENTIFIC MEETINGS slope of the concentrationequation for CX,Y,Z; e is the base of natural logarithms. The unknown parameters are estimated by well-knownmethods of analytic geometry. First ro is estimated by linearizingthe equation for Cxyrrwhich then allows K to be estimated by linear regression. Finally eff is estimated by min- imizing the Chi Square statisticon the observed versus the ex- pected PV difference, Chi Square= (dVo - dVz)‘/dVz. These equations can then be appliedto biological data col- lectedin a few planes tangentialto the source such that a point and a discsource of ionscan bedistinguished. Suchsources are very often quite weak, so a simple one-dimensional transect away from a source does not yield much data. Literature Cited 1. Kiihtreiber, W. M., and L. F. Jaffe. 1990. .I. Cell Biol. 110: 1565- 1573. 2. Smith P. J. S., R. H. Sanger, and L. F. Jaffe. 1994. Meth. Cell Biol. 40: 115-134. Reference: Biol. Bull. 187: 272-273. (October, 1994) Lobster Orientation in Turbulent Odor Plumes: Simultaneous Measurement of Tracking Behavior and Temporal Odor Patterns Jennifer Basil and Jelle Atema (Boston University Marine Program, Marine Biological Laboratory) Chemical cues play an important role in the food searching behaviorof the lobster, Homarus americanus. Previous studies have shown that bilateral antennular chemoreception isnecessary for search efficiency within a turbulent odor plume (1, 2). The manner in which lobsters orient within plumesindicates that thereis directionalinformation contained within turbulent odor plumes (3). In addition, using chemical sensors (In Vivo Electrochemical System, IVEC) capable of measuring a tracer (dopamine) in sea- water, Moore and Atema (4) measured the fine-scale, three-di- mensional structureof plumes. Plumes showed spatial gradients of suchfeatures as pulse height, onsetslope, and distribution, which could provide directional cues to orienting animals. The purpose of the present study was to implementtechniques from which wecould determine whether animals extract anduse fine- scale features to successfully navigateto a distant odor source. In the presentexperiment, behavioral and electrochemical measurements weremade in a flow-through flume (250 X 90 X 20 cm). The food stimuluswas gravity-fed through a Pasteur pipette positionedso that the nozzle released stimulusat the cross-sectional center of the tank, 9 cm off the bottom, at the upstream end of the flume. The lobstercarriedtwo IVEC elec- trodes mounted directly over the lateral antennules, and a sub- mersible amplifier on its back.The backpack was connected, by a 2.5-mflexiblecable, to a computer that ran the electrochemical software (IVEC). Theoverhead film recordof the freelyorienting lobster could be synchronized with the real-timeconcentration measurements. Lobsters werefed twice weekly and weredeprivedof food for 3- 10days prior to eachtrial (3). Twenty-four hours before a trial, a lobster was fitted with a blackplastic blindfold andallowed to habituate to the orientationarena. Food extracts (squid, clam, mussels) with tracer were used as stimuli (source tracer concen- tration = 40 mA4). The stimulus wasintroduced into the arenawhen a lobster settled into the downstream shelter. Localizationwas considered successful only if the lobster approached to within 15 cm of the . Figure 1. Onset slope and heading angle, plotted together as a function of time (t = 0 s at start of orientation trial, t = 30 s when animal located source). Onset slope is plotted continuously at 10 Hz. Onset slope from electrode mounted over right antennule is top trace (solid line). Onset slopefrom left antennular area is plotted as the lower trace (dashed line). Large onset slopes are located close to the source, where heading angle improves. pipettetip within 20 min after the start of stimulus introduction and did not walk alongthe side wallsof the flume. Orientation paths were digitized at l/s. Variables chosen to quantify tracking behavior includedwalking speed, turning angles, and headings. Backpacks did not affectorientation behavior.Figure 1illus- trates the onsetslopes for odor tracer patches encountered by the left and right electrode (right electrode = positive values, left electrode = negativevalues),the slopes are plotted abovethe lobster’s heading as it oriented towardsthe source (lobster hits sourceat 30 s). Positive headings indicate that the animal is moving to the right of the line connecting the starting point and the source. The odor patches, as encountered by the moving lobster, closely resemble those measured in previous studies (4). At points in the plume that are distant from the odor source, there werea greaternumber of patches of lower concentration and shallow onset slopes. Close to the source, the patches areof