Dietary Si (orthosilicic acid; OSA) appears important in connective tissue health, and although the sources and intakes of Si are well established, its absorption is not. Si absorption was measured from eight high-Si-containing sources: alcohol-free beer; OSA solution (positive control); bananas; green beans; supplemental choline-stabilised OSA (ChOSA); supplemental monomethyl silanetriol (MMST); supplemental colloidal silica (CS); magnesium trisilicate British Pharmacopoeia antacid (MTBP). Two of the supplements and the antacid were pre-selected following an in vitro dissolution assay. Fasting, healthy subjects (CS, n 3; others, n ≥ 5) each ingested two of the sources separated by a 1-week wash-out period. Blood and urine were collected and measured for total Si concentrations by inductively coupled plasma optical emission spectrometry. Absorption, based on urinary Si excretion, was highest for MMST and alcohol-free beer (64 % of dose), followed by green beans (44 %), OSA (43 %), ChOSA (17 %), bananas and MTBP (4 %) and CS (1 %). Peak serum concentrations occurred by 0.5 h for MMST and green beans, 1.5 h for OSA and alcohol-free beer, 2 h for ChOSA and CS, and 4 h for MTBP. Area under the serum curves correlated positively with urinary Si output (r 0·82; P < 0·0001). Absorption of Si from supplements and antacids was consistent with their known chemical speciation and kinetics of dissolution under simulated gastrointestinal conditions. Monomeric silicates were readily absorbed, while particulate silicates were decreasingly well absorbed with increasing polymerisation. The present results highlight the need to allow for relative absorption of Si from different foods or supplements in subsequent epidemiological and intervention studies.