Fig. 5
Worms improve their orientation towards the target following the exit from a pirouette. a (I) Angular bearing histogram of worms oriented off-course (90° < B < 270°) immediately before initiating the pirouettes (B Before ); (II) Angular bearing histogram of the worms from (I) immediately after the pirouettes (B After ); (III) A histogram of the cosine of the difference in the bearings before and after a pirouette cos(B Before – B After ) for worms that were initially off-course. The histogram shows two peaks (around 1 and –1) indicating that worms tend to perform either extreme (e.g., 180°) or minute (e.g., 0°) angle changes. However, the tendency to perform a pirouette with a larger angular difference in bearing is significantly more probable (P < 10–6, permutation tests, see Methods). The data is composed of 13,297 disoriented pirouette events. b (I) Angular histogram of bearings for worms oriented on-course (0° < B < 90° or −90° < B < 0°) immediately before the initiation of a pirouette (B Before ). (II) Angular histogram of worms bearing that are on-course immediately after the pirouette (B After ). (III) A histogram of the difference in the bearings before and after a pirouette cos(B Before – B After ) for on-course worms. As in off-course worms, these initially on-course oriented worms tend to perform either extreme or minute angle changes, but the frequency of minute changes (cos(Δ angle) = 1) is significantly higher (P < 10–6, see Methods). The data is composed of 15,368 oriented pirouette events. c Simulations of chemotaxis trajectories. We used three different strategies for choosing the exit angle from a pirouette (see Methods for details). The experimentally observed principle, where the exit angle is sampled according to the entry angle, provides an efficient chemotaxis strategy reflected by the significantly shorter time to reach the target point (P < 0.007 and P < 2.5 × 10–3, Wilcoxon rank-sum test for random and uniform sampling, respectively). Error bars denote SEM of the number of simulated worms in each simulation. Overall, we simulated 250 worms per strategy