The mechanisms and neurological pathways governing pupil size and reactivity are complex and still not fully understood in brain damage and disease [l]. Until very recently ‘widely dilated fixed pupils’ were generally accepted as a reliable sign of irreversible brain damage and a strong justilication for abandoning unsuccessful cardiopulmonary resuscitation (CPR), despite the wide variation in pupil size in this context and their longrecognised unreliability as a guide to cerebral oxygenation and integrity, including brain death [2-41. Whilst a reduction in pupil size is one of the most valuable signs of effective CPR, and often more forceful chest compression can delay pupil enlargement in protracted resuscitation, so far as I know there has been no clinical study on the relationship between pupil size and cerebral blood flow in this situation. A change which is surely long overdue is the final abandonment of still widely used terms (even in the literature) to describe pupil size, such as ‘small’, ‘moderately constricted’, ‘widely dilated’ and so on in favour of accurate assessment in millimetres [4], as has long been standard practice with the Glasgow Coma Score [5]. As to ‘widely dilated’ (2 8 mm diameter) pupils being ‘fixed’, do we ever see pupils of this size in this situation react to light, since the reflex seems to be abolished by severe cerebral ischemia, and reappears only as the pupils return to normal size (~6 mm) and then not invariably in CPR? My personal experience suggests that as in general anaesthesia a positive eyelash reflex during CPR signifies that consciousness is only slightly depressed. It also suggests the brain is intact and the patient will swiftly regain consciousness once satisfactory heart action returns [4]. However, it generally disappears when CPR lasts for more than several minutes. The light reflex is a valuable sign in trauma patients [l] but how useful and how reliable a guide to brain status is it in CPR, and how does it compare with monitoring the eyelash reflex and pupil size? The concept of a protective penumbra during brain ischemia, between 10 and 30% of normal cerebral blood flow (CBF), structural integrity with functional silence [6], offers valuable insights into the possible impact of CPR on the brain [7]. It is likely that each part of the brain has its own critical flow value, highest in selectively vulnerable areas, in particular the cerebral cortex, its filter the reticular formation of the thalamus and the hippocampus [8-lo]. It is also likely the ischemic penumbra is, at least initially, flow time-dependant [l 11, its value rising as CPR continues (Fig. 1). One large study [ 121 has shown that provided CPR starts within about 5-6 min of cardiac arrest and