Fig. 1: C. elegans’ response to bacterial density follows a universal sigmoidal trend.

a Experimental scheme: Worms are placed at the center of a regular pentagon formed by 5 food patches of different densities. After 2 h of exploration, worms located at each food patch are counted. b Relative number of worms found at each food patch, as a function of bacterial density in the food patch (D). Squares: Experimental data for E. coli OP50. Line: Fitted sigmoid, following Eq. 1 in Methods. c Sigmoid parameters: \(H\) is the ratio between the number of worms at the high and low density extremes, \(k\) is the slope at the sigmoid’s midpoint, and \({D}_{{{\mbox{attract}}}}\) is the density at which the number of worms reaches 5-fold the low-density baseline. d Same as (b), but for all bacterial strains and without errorbars (see Supplementary Fig. 4 for separate plots for each strain and Supplementary Fig. 5 for the parameters of all sigmoids) e Measured proportion of worms in each food patch, versus proportion predicted by the sigmoid, fitted to each strains (Eqs. 1 and 2). f Relative number of worms found at each food patch, as a function of effective density (\({D/D}_{{attract}}\)). Black line: Sigmoid with \(H=146\), \(k=1.4\). g Same as (e), but with predictions made using effective density and the same sigmoid for all strains. All errorbars show the 95% confidence interval, computed via bootstrapping; see Supplementary Table 1 for sample sizes.