Human physiological finger tremor is described as an inevitable small trembling of the finger consisting of two main frequency components; one 8-12 Hz and one over 15 Hz. When the finger is extended, both of these peaks are present in the frequency spectrum, but their relative size is variable. When moving the finger, the low frequency peak is much larger and usually the high frequency peak is no longer identifiable. Our recent studies revealed a main role for mechanical resonance in generating finger tremor. In contradiction to previous beliefs, a specific neural input was not necessary. Both tremor frequencies could be explained based on a changing resonant frequency due to altered muscular stiffness with movement. How the tremor frequencies transform during the transition between posture and movement and the interaction with EMG is described here. With ethical permission, physiological tremor of 15 healthy subjects (23.7 ± 9.9 yr, 12 male) was measured using a miniature accelerometer taped to the splinted middle finger. A retroflective laser sensor, pointed at the tip of the finger, measured finger position. A pc placed ~2 m in front of the subjects showed finger position and a target with which subjects had to align their finger position. The target started with 10 s at a comfortable static middle position of the finger which progressed into a vertically orientated sinusoidal chirp signal ranging from 0 Hz to 0.05 Hz over 50 s. Over the subsequent 60 s the target consisted of a mirror image of the first half, ending with a 10 s static middle position (see Fig 1a). The maximum required angular velocity of the finger was 3 deg/s. Each subject was asked to repeat the 120 s trial 10 times, during which surface EMG was recorded from the extensor digitorum communis muscle (m. EDC). We calculated a wavelet transform for acceleration and rectified EMG for each trial which afterwards were averaged across trials and participants. A Pearson’s correlation was computed between acceleration size and EMG size and between these measured variables and position and speed of the finger. Unsurprisingly, EMG correlated strongly with the vertical position of the finger (r^2 = 0.84) (Fig 1, top), because more extension required more EMG. In contrast, acceleration size seemed to correlate best with the speed of the finger (r^2 = 0.72) (Fig 1, bottom), and less strongly with EMG (r^2 = 0.19). We propose that primarily the movement-related (thixotropic) character of the muscles determines the size of physiological finger tremor. With increased speed of finger movement, an increased amount of muscle tissue is moved about and the overall muscle stiffness will progressively drop. This is associated with an increase in the size of the resonance and thus finger tremor. This must cause computational difficulties in finger control when making the transition from posture to movement.