Trabecular bone disconnection is an independent factor in age-related skeletal failure where real termini (ReTm; rare in youth) may cause weakness disproportionate to tissue loss, yet their structural contribution at vulnerable locations remains uncertain. ReTm (previously recorded at the iliac crest) were mapped in "normal" aged vertebral bodies (T11-L5 autopsy; 20 females, 10 males) and corresponding proximal femora (autopsy; 10 females). Results were compared with biomechanically failed femora from orthopaedic subjects aged >58 yr (osteoporosis OP, 10 females; osteoarthritis OA, 10 females). A novel direct 2D/3D histological method was applied to large, thick (300 μm) slices superficially silver-stained to separate ReTm (unstained) from apparent termini (planar artefacts, brown). Light microscope field co-ordinates enabled ReTm mapping and statistical testing relative to i) sex, ii) tissue sector and iii) slicing plane. In men ReTm populations were small and random while in women they were large and sector-specific. In vertebrae they clustered anterior/superior being rare posterior/inferior; in the femoral head they concentrated distal/superior and also near the fovea, being fewer distal/inferior. A distribution polarity was evident with 100% more ReTm observed transversely (i.e., on tensile-related cross struts) than longitudinally (i.e., on compression-related vertical struts). Their numbers rose in OP (BV/TV14%), remaining polarised and sector-specific in OP only. Comparative experimentation by marrow elution of an OP animal model demonstrated "floating segments" as a possible outcome. Conclusions were supported statistically that trabecular disconnection "hotspots" at vulnerable locations are sex- and sector-specific, mainly transaxial, and subject to disease modulation.