Abstract
Conditions for local instability are compared with conditions for global asymptotic stability of an output error adaptive estimation algorithm for continuous-time systems. A boundedness conjecture is made that all signals within this adaptive system including the output error and parameter estimates remain globally bounded for all time despite any locally unstable behaviour. A global stability criterion applied to a transfer function, which is not strictly positive real, states that for a large enough adaptation gain the system will be asymptotically stable. However, a local result states that for a small enough adaptation gain the same system is unstable. The results suggested by the computer simulations are verified by an exact analysis of a linearized periodic version of the adaptive system.