Fig. 4. Starvation and sustained exposure to new temperatures have separate long-term effects on thermotactic behavior. (A) When wild-type (WT) worms that were grown overnight at 15°C and then starved for 4 h at 25°C are placed near the middle of the linear thermal gradient, they exhibit random dispersal. The movements of (B) wild-type worms and (C) ttx-3(ks5) mutant worms up or down spatial thermal gradients spanning 18–22°C were quantified after they had been grown at 15°C (left-hand panels) or grown at 25°C (right-hand panels). Line colors indicate the temperature of the worms during the 4 h preceding each experiment; blue represents 15°C and red represents 25°C. Solid lines indicate experiments using unstarved worms. Dotted lines indicate experiments in which worms were starved at 15°C or 25°C before each experiment. In each data trace, error bars (±1 s.e.m.) are shown at 10 min intervals. The speed of instantaneous forward-crawling movements exhibited by individual worms in each experiment is indicated in italics, showing that atactic behavior cannot simply be attributed to lack of mobility (mean ± s.d.).