In order to assess the risks to human health posed by elevated concentrations of arsenic in vegetables, and to identify pollution-tolerant vegetable varieties, a large scale survey of arsenic levels in soils and vegetables planted or sold in Beijing was conducted. Thirty-nine soil samples were collected from gardens and fields used to grow vegetable plants. In addition, 93 varieties of more than 400 fresh vegetable samples were obtained from vegetable stalls, supermarkets and wholesale outlets. Arsenic concentrations were measured using hydride generation atomic fluorescence spectroscopy (HG-AFS). Arsenic concentrations in soils ranged from 4.44 to 25.3 mg kg-1, with arithmetic and geometric means of 9.40 and 8.79 mg kg-1, respectively. Compared with the background arsenic concentrations of soil from Beijing, there appeared to be a significant accumulation of arsenic in soil collected from gardens/fields that produce vegetables. Arsenic concentrations in the edible plant portions ranged from less than the analytical detection limit (0.1 μg kg-1 fresh weight) to 0.479 mg kg-1 fresh weight, with a mean of 0.028 mg kg-1 fresh weight. In all of the samples, arsenic was less than the Tolerance Limit of Arsenic in Foods for China (TLAFC) of 0.5 mg kg-1 fresh weight. The TLAFC is the maximum permissible concentration of arsenic in vegetables that will be consumed by people. The highest level of arsenic detected in a vegetable plant was 0.479 mg kg-1, which was measured in a radish (Raphanus sp.) sample obtained from the Fengtai District of Beijing. Arsenic was detected at 0.331 mg kg-1 in a garlic sample collected from Shandong Province, which is higher than the standard of 0.25 mg kg-1 set by WHO/FAO. The spatial distribution of arsenic concentrations in vegetables planted in Beijing presented apparently a "U-shaped" pattern at the northeast deflection angle of 11o-30o. The arsenic concentration in field-grown vegetables was significantly higher than the concentration of those planted in a greenhouse.Results of hierarchical cluster analysis on the arsenic bioconcentration factor (BCF) in vegetables indicated that the plants sampled could be separated into two groups based on BCF. Rape (Brassica campestris), radish (Raphanus sp.), pakchoi (Brassica chinensis), onion (Allium fistulosum), mustard (Brassica juncea), cucumber (Cucumis sativus), Chinese cabbage (Brassica pekinensis) and cabbage (Brassica caulorapa) had higher arsenic BCFs while chili (Capsicum annuum), beans round trellis (Vigna unguiculata), wax gourd (Beninacasa hispida), eggplant (Solanum sp.), spinach (Spinacia oleracea), tomato (Lycopersicon esculentum) and celery (Apium graveolens) had lower arsenic BCFs. The average ingestion rate of arsenic from vegetables was 0.016 mg/person/day for people of Beijing. Consuming vegetables with elevated arsenic concentrations may pose a health risk to local residents, particularly the young, elderly, or ill.