280 JOURNAL OF COSMETIC SCIENCE % removal 8ø I 6O 4O 2O 100 no. oscillations 150 [] zero abr :'D 2% abr •'D 4% abr [] 10% abr 33% paste slurry 150 cpm oscillation rate Figure 1. Film removal as a function of abrasive silica content of paste. following six weeks' use. Corresponding British Standard Dentine Abrasion Values (DAV) (8) are also given. The stain-prevention performances of the 10% and 4% abrasive silica pastes were similar (Mann Whitney U test) and significantly more effective (p 0.001) than the 2% and zero-abrasive versions. The virtually identical stain-prevention action of the pastes con- taining 10% and 4% w/w abrasive silica is achieved despite a substantial difference in measured dentine aggression. Comparison of the data with that of Figure 1 indicates a remarkable degree of coinci- dence of effect. This is illustrated in Figure 2. DISCUSSION Formulators of abrasive toothpastes for general use should aim to provide the maximum Table I Six-Week Clinical Stain Formation as a Function of Abrasive Silica Content and Dentine Abrasion Potential of Four Pastes % Abrasion silica % Visible stainers Mean total-stain score (SE) DAV 10 19 4.2 (0.33) 55 4 19 4.7 (0.35) 30 2 32 9.5 (0.34) 17 0 65 11.2' 5 * Incomplete six-week score owing to exercise of cosmetic withdrawal option.

DENTAL STAIN PREVENTION 281 % 80 60 40 20 0 I , I • I , I • I , I 0 2 4 6 8 10 % abrasive in paste FT model film removal (%) visible stainers (% of panel) [] Figure 2. Comparison of in vitro and clinical stain-prevention data for pastes of different abrasive content. stain-control function consistent with the minimum dentine abrasion potential. Most marketed toothpastes implicitly claim to have this balance despite containing a wide range of levels and types of abrasive. It is unlikely, however, that all demonstrate the optimum balance of care and efficiency, particularly in view of the widely held mecha- nistic hypothesis that the abrasive is included for stain removal. If, however, one recog- nizes that the primary realistic function of the abrasive in toothpastes is that of stain- prevention, it becomes apparent that different conditions may apply and that an in vitro test more closely simulating the removal characteristics of immature pellicle is required. The data above suggest a strong clinical predictive function for the in vitro FT test in relation to abrasive stain prevention within a normal domestic regime of one to two brushings a day. It thus becomes a valuable tool for abrasive screening and optimization within the science of dentifrice formulation. Toothpastes for general sale and use are formulated to satisfy the requirements of a majority of consumers, and thus the human panel used represented a typical range of age and personal habits. It became clear during clinical pilot studies that the rate and extent of stain formation following professional cleaning was strongly influenced by the sub- ject within subjects, however, behavior appeared relatively invariable. This is consistent with the belief that toothbrushing is a ritual activity, and that skill and diligence vary considerably between subjects. Long-term parallel test designs would require careful habit screening and balancing to obtain panels that are sensitive only to paste cleaning