Fig. 2: C. crescentus chemotaxis.

a Example trajectory showing alternating backward (red) and forward (blue) runs, separated by switches in flagellar rotation direction that result in reversals (CW to CCW, teal) or flicks (CCW to CW, orange). b Schematic of E. coli run-tumble motility. Runs are driven by CCW rotation (blue). Tumbles result from temporary CW rotation (red) of at least one flagellum and thus are bordered by two switches in rotation direction (CCW to CW, teal, CW to CCW, orange). c Radial plot of average run durations (blue: forward, red: backward runs) as a function of projected angle to the x-axis in the x-y plane. The dashed lines indicate the average run durations observed for swimming down the gradient and serve to facilitate comparison. Shading indicates 95% confidence intervals. Values are computed from a total of 2898 forward and 5342 backward runs. d Schematic of two-state motor rotation model proposed for C. crescentus and e established for E. coli^{38, 39}. The energy difference ΔE_b (purple) between states determines the motor bias, while the energy barrier between states, ΔE_ccw and ΔE_cw, determines the switching rates, k_cw and k_ccw, respectively (see Supplementary Discussion). f Average forward versus backward run durations, <τ_fw > versus <τ_bw > , up (yellow), down (cyan), or perpendicular to (green) the gradient (defined by 36° cones around positive x-axis, negative x-axis, or y-axis, respectively). The solid line reflects a best-fit constant CW bias of 0.54 ± 0.01, with the standard error (Methods) shown as gray shading. P values are shown for one-sided t-tests between durations up and down the gradient. Error bars reflect standard error of the mean. Averages and p values are determined over 506, 449, 408 forward runs and over 882, 946, 835 backward runs ascending (yellow), descending (cyan), and perpendicular to (green) the gradient, respectively.