All values represent mean ± SE. Significant pairwise comparisons (all two tailed and Bonferroni corrected) for accuracy are as follows: textures in experiment 2, shared vs different, t₍₁₃₎ = 3.83, p = 0.006; textures in experiment 3, shared vs different, t₍₁₃₎ = 4.53, p = 0.002. Significant pairwise comparisons for response latency are as follows: object ensembles in experiment 2, shared vs identical, t₍₁₃₎ = 5.05, p = 0.001; textures in experiment 2, shared vs identical, t₍₁₃₎ = 4.06, p = 0.004; shared vs different, t₍₁₃₎ = 2.86, p = 0.036; object ensembles in experiment 3, shared vs identical, t₍₁₃₎ = 5.36, p = 0.001; shared vs different, t₍₁₃₎ = 3.52, p = 0.011; textures in experiment 3, shared vs identical, t₍₁₃₎ = 3.01, p = 0.03; shared vs different, t₍₁₃₎ = 3.34, p = 0.017; object ensembles in experiment 4, shared vs identical, t₍₁₀₎ = 3.89, p = 0.01; shared vs different, t₍₁₀₎ = 3.49, p = 0.017.