Twenty patients (12 men and 8 women, median age 69 years) with metastatic tumours in the parotid gland who presented over a 12-year period were evaluated retrospectively. Preoperative investigations included fine needle aspiration cytology (n = 11) and computed tomography or magnetic resonance imaging (MRI) (n = 14). Most tumours originated from the head and neck region, the two main types being squamous cell carcinoma (n = 10) and malignant melanoma (n = 7). All 20 presented with a parotid mass and 11/20 (55%) had associated lymphadenopathy. Eleven patients (55%) underwent superficial, five total, and four radical, parotidectomy. Neck dissection was required in 16 patients (80%), and all 11 patients with clinically palpable lymph nodes had evidence of tumour in the neck dissection specimens. Half of all patients (n = 10) received adjuvant postoperative radiotherapy. Three-quarters of the patients (n = 15) were alive after a mean follow-up of 31 months and only one developed a marginal recurrence. The cumulative 5-year survival rate was 51%, and there was no significant difference (P = 0.48) in the 3-year survival rates of patients who had radical compared with those who had modified neck dissections. Patients who had superficial parotidectomy had a longer overall survival compared with those who had total or radical parotidectomy (P = 0.04) perhaps reflecting the advanced nature of tumours that required total or radical excision of the gland. We conclude that superficial parotidectomy is usually an adequate treatment for secondary parotid tumours (when disease is clinically limited to the superficial lobe), and we suggest that patients in whom metastatic disease of the parotid gland is suspected do not require neck dissection if they have no palpable lymph nodes and MRI shows no evidence of spread. There seems to be no survival advantage in radical over modified neck dissection.