It has been known for many years that there is a Horner-like picture on the symptomatic side in many cases of cluster headache (1, 2). The corneal indentation pulse (CIP) amplitudes and the intraocular pressure (IOP) increase markedly during attacks, the increase being most pronounced on the symptomatic side (3, 4). As far as we know at present, these changes occur in all cases, provided the attack is pronounced enough. CIP amplitudes and lOP were the first cluster headache parameters which were demonstrated to show attack-related changes without exception. The corneal temperature also increases on the symptomatic side during attacks (4). Interictally, there may be "cold spots" supraorbitally on the symptomatic side as demonstrated by thermography (5), and they frequently transform into warmer areas during attacks (6). How studies provide contradictory results, but recent investigations (Doppler studies) tend to show that blood flow velocity is increased in the symptomatic area during attacks (7). Recently, nasal secretion and lacrimation have been quantified and have been found to increase bilaterally during attacks, although mostly to the largest extent on the symptomatic side (8, 9). All these features are presumably reflections of an autonomic disorder which is most pronounced on the symptomatic side.What do these "autonomic" disturbances imply as far as the underlying mechanism is concerned?