Purpose: Investigation of the influence of radiation dose to salivary glands on xerostomia in patients with head and neck cancers. Methods: From October 2002 to December 2011, 548 patients with head and neck carcinomas were treated in our department using intensity modulated radiotherapy (IMRT). 325 patients were eligible for analysis more than 1 year after irradiation. Xerostomia was evaluated according to the criteria of Radiation Therapy and Oncology Group (RTOG) and xerostomia related questionnaire. For statistical analysis grade 1 and 2 were recorded as mild and grade 3 and 4 as severe xerostomia. The dosimetric values for absolute mean dose, biological equivalent mean dose (EQD2), the relative organ volume not exceeding 20 Gy, 25 Gy and 30 Gy (V20, V25, V30) for all 4 major salivary glands or for both parotid glands only were used for analysis. V20 revealed the best discrimination between both patient groups (with vs. without xerostomia), compared to parameters V25 and V30. Therefore the volume of the salivary glands receiving less than 20 Gy (V < 20Gy) was analyzed additionally. Mann-Whitney-U-test, Kruskal-Wallis-test and logistic regression were used in statistical analysis. Results: A TD 50 can be determined for the occurrence of xerostomia more than one year after radiotherapy for both parotid glands of 19.3 Gy for the mean dose absolute, 11.2 Gy for the mean dose EQD2, 38.2% for V20, 26.8% for V25, 18.7% for V30, 34.9 ml for V < 20Gy. For all major salivary glands theses values were 25.7 Gy for the mean dose absolute, 15.6 Gy for the mean dose EQD2, 51.3% for V20, 41.1% for V25, 33.9% for V30 and 34.8 ml for V < 20Gy. Conclusions: The identification of a tolerance dose for the salivary glands for treatment planning appears to be difficult, as the dose-response correlation only shows a flat slope. Additionally, a large interindividual variability seems to exist. We could not found any threshold dose for development of xerostomia.