Is coronary spasm the explanation for the paroxysmal myocardial ischaemia of Prinzmetal angina? Although considered adventurous two or three years ago, this hypothesis has been supported by the publication of approximately 15 cases in which the cause and effect relationship seemed to be proven by spasm seen on coronary arteriography at the same time as a typical attack. The authors report 13 cases of this type (the first of which was published in 1972). Spasm occurred in a normal coronary system in 3 cases, whilst in 2 cases there were only non-stenotic plaques of atheroma. In 8 cases, spasm occurred at the site of or distant from a significant atheromatous lesion. In 8 cases, spasme involved the right coronary whilst in 3 cases the A.I.V. was affected and in 2, two of the three main coronary trunks. In the light of their experience, the authors stress two points: --in many cases, nitroglycerine alone is not sufficient to relieve the spasm, and only the in situ injection of papaverine had a constant effect; --it is essential to opacify both coronary systems during a typical attack in order to obtain proof of spasm.