One aim of the Vg study of headache epidemiology was to depict the total panorama of headaches in a small-sized, Norwegian community at the end of the 20th century. In the present part of the study, a search was made for the rare, global headaches. Various, relatively unknown headache categories were observed, such as hydrogen sulphide intoxication (n = 2)-after exposure to H2S ambient air concentrations of ≥100 p.p.m. In Vg, nitroglycerine headache (‘dynamite headache‘) proved to be a relatively frequent disorder (n = 13). This represents approximately 0.7%of the study group. A main reason for this, in all probability unusually high frequency is that there previously have been soapstone mines within the precincts of the parish. Another relatively frequently occurring headache followed exposure to wind (n = 7; 0.38%). Ice-cream headache was rare (n = 3; 0.16%), when only the general question was asked: ‘other headaches?’. Approximately 7 years later, an ad hoc procedure was adopted: 50 parishioners who did not answer positively concerning ice-cream headache on the first examination were asked specific questions concerning ice-cream headache. Four had had such an experience, i.e. prevalence of 8%, which is 50 times higher than the original result, but still a rather low prevalence when compared with the mean prevalence from other studies: 39%. This demonstrates that the result as regards prevalence, even for a headache such as ice-cream headache, to a large degree depends upon the interview technique used. One aim of the Vg study of headache epidemiology was to depict the total panorama of headaches in a small-sized, Norwegian community at the end of the 20th century. In the present part of the study, a search was made for the rare, global headaches. Various, previously unknown headache categories were observed, such as hydrogen sulphide intoxication (n = 2) -after exposure to H2S ambient air concentrations of ≧ 100 ppm In Vg, nitroglycerine headache (’dynamite headache’) proved to be relatively frequent A main reason for this, in all probability unusually high frequency is that there has previously been been soapstone mines within the precincts of the parish. Another relatively busy occurring headache followed Approximately 7 years later, an ad hoc procedure (when n = 3; 0.16%), when only the general question was asked: ’other headaches?’ was a dopted: 50 parishioners who did not been related to ice-cream headache on the first examination were asked specific topic concerning ice-cream headache. Four had had such an experience, ie prevalence of 8%, which is 50 times higher than the original result , but still a rather low prevalence when compared with the mean prevalence from other studies: 39%. This demonstrates that the result as perceived prevalence, even for a headache such as ice-cream headache, to a large degree depends upon the interview technique used .