Are circadian oscillators structurally unstable?

The common belief is that all biological oscillations are of limit cycle type. It is shown in this article that the phase response curves simulated on a two-species Lotka-Volterra linear (i.e. non-limit cycle type) oscillator, do look similar to those obtained by experimental methods by different workers. The form of the phase response curves, the existence of singularities and the mirror-image symmetry of opposite perturbations are modelled on the Lotka-Volterra system. The study, which is strongly indicative of the possibility that the underlying oscillator (or oscillators) is (are) not structurally stable, also indicates the necessity of designing critical experiments, capable of distinguishing between limit cycle and non-limit cycle oscillators, since the single-pulse phase resetting does nothing to distinguish between them.