Pubmed shows that the annual total of papers mentioning physiological tremor has grown steadily from ~ 200 to ~ 400 over the last twenty years. One particular interest is the idea that a centrally generated, approximately 10 Hz component plays an important role in tremor generation. The concept of an autonomous central “tremor generator” is nearly as old as the discovery of the alpha rhythm. The discovery of a brain rhythm which had about the same frequency as tremor lead quickly to the idea that it “must” be the cause. This was quickly shown to be false, but the idea of a central generator is persistent. Work by Elble and colleagues established the central tremor generator (“an oscillating system of neurones located within the CNS”, Elble and Koller (1990)). It is now common to divide tremor into “central” and “mechano-reflex” components. More recently, the central oscillator theory has been invoked to explain the approximately 10 Hz tremor-like discontinuities that can be observed during movement (for example, Vallbo and Wessberg (1993)). This has suggested that all slow movements are executed on a pulsatile basis. The idea of obligatory pulsatile control is actually quite old (Travis 1929, Tiffin and Westhafer, 1940)) but has recently been reworked (Bye and Neilsen, 2010; the BUMP theory). There are at least two difficulties with the idea of a central generator. First, how many such generators are there? One per motor unit, one per muscle, one per movement or one per person? Second, central generators operating at ~ 10 Hz during posture and movement have not been confidently identified. The other half of the tremor story (the “mechano-reflex component”) has also been studied. The “mechano” part is the resonant properties of the limbs. It may be augmented by appropriately phased “reflex” activity (Lippold, 1970). Resonance results from the interaction of the stiffness of the muscles and the inertia of the moving parts of the limb. In recent work, we have shown that the complete spectrum of tremor can be economically explained by resonance. A simple way to demonstrate this is to see if a specific form of EMG is necessary to generate tremor, or whether a realistic tremor can be created by an entirely random neural input. In one approach we compared real tremor to that produced by a simple resonant model. The resonant model recreated the tremor spectrum when driven by real EMG or by an entirely random artificial input. We concluded that the contribution of any central component must therefore be minimal (Lakie et al 2012). In the second approach we drove relaxed limbs with random artificial excitation and again showed that a realistic tremor resulted. One of the key planks of the division of tremor into central and mechano-reflex components was that they occupied distinct and identifiable parts of the frequency spectrum. For some time, however, it has been known that the resonant frequency of a limb cannot be described by a single value. Skeletal muscles are thixotropic so that they decrease their stiffness with movement (Lakie et al 1984). We have shown that this feature can also be captured very well by a simple resonant model. Fig 1 shows hand tremor acceleration spectra recorded from a postural (static) and slowly moving (dynamic) limb. During movement the tremor size greatly increases (by approximately an order of magnitude) and decreases in frequency. This behaviour is also recreated by our simple resonant model using appropriately sized random inputs or recorded EMG. For finger tremor the resonant frequency is thought to be higher than the wrist – usually greater that 20 Hz. Recent work leads us to question this assumption, because the frequency of finger resonance will reduce very considerably with movement. The tremor spectrum with its two peaks is usually thought to represent coexisting and distinguishable central and mechano-reflex mechanisms. We alternatively suggest that the two peaks may both represent resonance. The high frequency peak occurs when the finger is stationary and the muscles are stiff and the lower frequency occurs during movement when the muscle stiffness is much decreased. We suggest that the finger tremor frequency is related to the speed of movement. When quasi-stationary, (postural), stiffness and resonant frequency are high and during movement both are reduced. Our conclusion is that as far as physiological tremor is concerned the role of a central component is negligible. Peripheral factors are extremely important and mechanical resonance is fundamental. It is possible that the resonant properties are complemented by reflex activity. One caveat is that the situation for pathological tremor is clearly different. Central oscillators are clearly involved in the genesis of Parkinsonian tremor and at least some forms of essential tremor.