The loss of muscle mass and strength with advancing old age is associated with, and possibly preceded by, loss of motor neurons and remodelling of the surviving motor units (Piasecki et al, 2015). The suggestion that tibialis anterior motor unit numbers can be preserved into old age by maintaining regular and intense exercise (Power et al., 2010) is appealing, but it remains to be shown whether any other muscles experience similar benefits. To investigate this, intramuscular electromyography (iEMG) and surface EMG (sEMG) were used to estimate motor unit numbers in the vastus lateralis (VL) of 21 young men (mean age 25±5y), 20 healthy older men (mean age 71±6y) and 12 male masters athletes who had trained and competed for most of their adult lives (MA; mean age 70±5y). The proximal and distal motor points of the VL were identified by percutaneous electrical stimulation. Surface EMG was recorded over each motor point and iEMG was simultaneously recorded adjacent to the sEMG in 12 locations of different depths within the muscle during isometric knee extension held at 25% of maximal voluntary contraction. A compound muscle action potential (CMAP) was generated by maximal percutaneous stimulation of the femoral nerve. Decomposition-enhanced spike-triggered averaging was used to obtain an average MU potential (MUP) area, and a corresponding surface-based MUP (sMUP) area. A motor unit number estimate (MUNE) was calculated at each motor point based on the CMAP divided by the average sMUP. A further estimate of motor unit numbers (intramuscular MUNE; iMUNE) was made by dividing the VL cross sectional area (CSA), measured from magnetic resonance imaging (Figure 1) by the mean MUP size. The results show that the young had the largest VL, with no difference between the Old and MA. Compared with the young, the Old and MA had 29% (p = 0.026) and 36% (p = 0.011) lower MUNE values, respectively, and 46% (p < 0.0005) and 50% (p < 0.0005), respectively, lower iMUNE values compared to the young. The Old and MA did not differ in either of the estimates of motor unit numbers. These results confirm the substantial neuromuscular deterioration that occurs in VL during healthy ageing and indicates that high levels of life-long exercise do not prevent the decline.