Two methods have been employed to measure the amount of tongue biofilm in humans: (1) areal density of microbes (cfu cm(-2)) (tongue scrape from a known/measured area), or (2) a tongue-coating index, where a human judge assesses the amount of 'coating' by visual inspection. The two methods were compared, both on human volunteers and on artificial biofilms. Fusobacterium nucleatum (Fn), Prevotella intermedia (Pi), Streptococcus mutans (Sm) and Peptostreptococcus micros (Pm) were obtained from stock cultures. Tongue scrape samples were obtained from four human volunteers in order to inoculate the in vitro matrices. A further 16 individuals (total n = 20) were co-measured for both visible tongue coating and quantitative biofilm (cfu cm(-2)). Tongue scrape suspensions were used to inoculate a cellulose matrix resting on a BHI-blood agar plate, periodically hydrated with sterile BHI and incubated for 10 days. Visual coating density images were recorded by camera. The relative light absorbance of standard suspensions of cells (pure species) was obtained by spectrophotometry across the visible spectrum. Biomatrix yields (log cfu cm(-2) ± SD) for Fn, Pi, Sm and Pm were 7.07 (±0.04), 7.38 (±0.07), 7.25 (±0.10) and 7.41 (±0.011). Higher yields were obtained for tongue-scrape mixed cultures, ranging from 7.47 to 7.76. Biofilms (in vitro and in vivo) had different appearances when measured by human judges in terms of colour, pigmentation and opacity. There was no correlation (R(2) < 0.02) between the two methods. Biofilm matrices of similar population densities (ca ∼ log 7.5) had widely different opacities depending on the species present.