Collection and analysis of DNA, most commonly from blood or buccal cells, is becoming more common in epidemiologic studies. Buccal samples, which are painless to take and relatively easily collected, are often the preferred source. There are several buccal cell collection methods: swabs, brushes, mouthwash, and treated cards, such as FTA or IsoCode cards. Few studies have systematically compared methods of buccal cell collection with respect to DNA yield and amplification success under similar conditions. We compared buccal DNA collection and amplification using buccal swabs and FTA cards in 122 control subjects from our Australian case-control study of childhood acute lymphoblastic leukaemia. Buccal DNA was quantified using a real-time PCR for beta-actin and genotyped at the loci of three polymorphisms (MTHFR 677C>T, ACE I/D, and XPD 1012G>A). PCR was successful with DNA from buccal swabs for 62% to 89% of subjects and from FTA cards for 83% to 100% of subjects, depending on the locus. The matched pair odds ratios (95% confidence interval) comparing success of FTA cards with buccal swabs are as follows: MTHFR 677C>T using PCR-RFLP, 12.5 (11.6-13.5) and using real-time PCR, 130.0 (113.1-152.8); ACE I/D using PCR-amplified fragment length polymorphism, 3.36 (3.2-3.5); XPD 1012G>A using real-time PCR, 150.0 (132.7-172.3). FTA cards are a robust DNA collection method and generally produce DNA suitable for PCR more reliably than buccal swabs. There are, however, technical challenges in handling discs punched from FTA cards that intending users should be aware of.