Table 1: Results from ANOVA F-tests for all 15 second-order variables of texture analysis (main effects and interactions).

Texture analysis parameterMain effects (P value)Effect interactions (P value)
(1) Exp. day(2) Potency(3) Replicate1-21-32-31-2-3

Cluster_prominence<0.0001 * 0.04360.21430.00050.11350.31260.1706
Cluster_shade<0.0001 * 0.0004 * 0.12960.32640.21450.43650.5069
Correlation<0.0001 * 0.00390.53760.0001 * 0.15670.09180.1896
Diagonal_moment<0.0001 * 0.0002 * 0.05080.29980.09430.61760.3498
Difference_energy<0.0001 * 0.00760.39480.00400.24440.16250.1077
Difference_entropy<0.0001 * 0.00590.50620.00090.22480.12880.1592
Energy<0.0001 * 0.00550.42460.01430.30510.16840.1437
Entropy<0.0001 * 0.01500.29970.00060.19700.19470.1318
Inertia<0.0001 * 0.00410.52620.0001 * 0.15670.09780.1934
Inverse_different_moment<0.0001 * 0.00240.01660.0001 * 0.08930.21930.0138
Kappa<0.0001 * 0.00230.0112<0.0001 * 0.09230.25210.0124
Maximum_probability<0.0001 * 0.00220.05280.29300.19580.13420.0886
Sum_energy<0.0001 * 0.0002 * 0.41630.00460.09020.03970.3965
Sum_entropy<0.0001 * 0.00130.79320.0001 * 0.13800.03870.1780
Sum_variance<0.0001 * 0.00300.6470<0.0001 * 0.16220.05070.1598

Note. Independent experimental parameters were (1) experimental day {LBI: 1–7, BRAD: 8–15; n = 15}, (2) treatment {Stannum met. 30x or water 30x; n = 2}, and (3) internal treatment replicate {1–3; n = 3}. Each experimental parameter combination (statistical treatment cell) was assessed by 6 biocrystallogram replicates. Highly significant effects (P < 0.01) are shown in bold. Statistical tests significant at P < 0.05 after a Bonferroni-Holm correction are marked with an asterisk.