Table 2 Results of post hoc (ANOVA) comparisons of bone CSG properties and linear regression standardized residuals.

All ANOVAs are significant to P < 0.001. Bold indicates significant post hoc P values. All values of J were size-standardized. Run, endurance runners; FB, football players; Row, rowers; Neo, neolithic; BA, Bronze Age; IA, Iron Age; Med, Medieval period; SR Hum:Tib, standardized residuals of regression of humeral and tibial J.

Bone variableNeoBAIAMedRunFBRow
Tibial Imax/Imin
  BA0.959
  IA0.5231.0
  Med<0.0270.5451.0
  Run0.734<0.033<0.012<0.001
  FB1.01.00.9990.5880.595
  Row1.01.00.9990.4530.2591.0
  Control<0.0140.8231.01.0<0.0010.8970.750
Tibial J
  BA0.774
  IA1.01.0
  Med0.9451.01.0
  Run<0.001<0.001<0.001<0.001
  FB0.269<0.0030.166<0.0190.801
  Row1.00.6641.00.86<0.0010.701
  Control1.00.9991.01.0<0.001<0.0461.0
Left humeral J*
  BA0.980
  IA0.9730.709
  Med0.093<0.0240.078
  Run0.3110.1380.1960.170
  FB<0.008<0.004<0.0080.9550.115
  Row<0.026<0.013<0.020.7240.3100.689
  Control<0.001<0.001<0.0010.297<0.0010.160<0.038
Right humeral J
  BA1.0
  IA0.8730.633
  Med0.1080.128<0.024
  Run0.9940.9990.4580.225
  FB<0.05<0.036<0.0050.9490.224
  Row0.7680.8180.1430.4880.9880.765
  Control<0.001<0.001<0.0011.0<0.0050.8080.081
SR Hum:Tib J
  BA0.247
  IA1.00.654
  Med1.00.1810.992
  Run0.072<0.001<0.0370.145
  FB<0.003<0.001<0.002<0.0170.983
  Row1.00.4610.9931.00.5720.163
  Control<0.009<0.001<0.0090.0791.00.7040.550

*Kruskal-Wallis test was performed due to non-normality of distribution in one of the groups, with two-tailed Mann-Whitney post hoc comparisons.

†Small sample size (n = 6).