Fig. 4. Plot showing percentage of correct trials averaged across the three bats (F11,22=1.691, P=0.142). N=30 trials per data point (10 per bat).