The experimental challenges presented by the analysis of trace volatile organic compounds (VOCs) in exhaled breath with the objective of identifying reliable biomarkers are brought into focus. It is stressed that positive identification and accurate quantification of the VOCs are imperative if they are to be considered as discreet biomarkers. Breath sampling procedures are discussed and it is suggested that for accurate quantification on-line real time sampling and analysis is desirable. Whilst recognizing such real time analysis is not always possible and sample collection is often required, objective recognition of the pitfalls involved in this is essential. It is also emphasized that mouth-exhaled breath is always contaminated to some degree by orally-generated compounds and so, when possible, analysis of nose-exhaled breath should be performed. Some difficulties in breath analysis are mitigated by the choice of analytical instrumentation used, but no single instrument can provide solutions to all the analytical challenges. Analysis and interpretation of breath analysis data, however acquired, needs to be treated circumspectly. In particular, the excessive use of statistics to treat imperfect mass spectrometry/mobility spectra should be avoided, since it can result in unjustifiable conclusions. It is should be understood that recognition of combinations of VOCs in breath that, for example, apparently describe particular cancer states, will not be taken seriously until they are replicated in other laboratories and clinics. Finally, the inhibiting notion that single biomarkers of infection and disease will not be identified and utilized clinically should be dispelled by the exemplary and widely used single biomarkers NO and H 2 and now, as indicated by recent SIFT-MS results, HCN and perhaps pentane and acetic acid.Hopefully, these discoveries will provide encouragement to research workers to be more open-minded on this important and desirable issue.2