In vitro measurements of aerosol fine particle fraction (FPF) using particle-sizing apparatus (e.g. the twin impinger, multi-stage liquid impingers, cascade impactors) have a key role to play in the development of new pharmaceutical products and in quality control. However, use of in vitro methodology to attempt to predict lung deposition in vivo is of limited value due, in part, to the inability of current apparatus to mimic upper and lower airway anatomy satisfactorily. Estimates of FPF based on cut-off points ranging from 5-7 microns generally overestimate lung deposition as measured in vivo by gamma scintigraphy. We recommend that: 1. multistage apparatus (minimum five stages) be used to characterize particle size distribution adequately, over the range 0.5-5.0 microns; 2. where possible, measurements should be made at a range of rates and profiles of flow reflecting those likely to be generated using the inhalation device in clinical practice (including use by young and elderly patients with varying degrees of airflow obstruction); 3. encouragement should be given to the further development, standardization, and validation of apparatus with a 'throat' which more closely resembles the human oropharynx and larynx. Pharmacokinetic methods can give a good estimate of total, but not regional, lung deposition, with drugs which are either not absorbed via the gastrointestinal tract, or whose absorption can be blocked by co-administration of charcoal, thus avoiding confounding by absorption of drug substance deposited in the oropharynx and subsequently swallowed. Techniques which rely on evaluation of a timed fractional output of drug substance in the urine are susceptible to the inherent variability of rate of absorption across the respiratory epithelium. We recommend that consideration should be given to the further refinement and validation of PK methods which would more clearly identify the fractional dose deposited in the lung. Lung-imaging methodology, e.g. gamma scintigraphy, employing formulations radiolabelled with gamma-ray-emitting radionuclides such as 99mTc, can measure total lung deposition and oropharyngeal deposition, provided that the radiolabelling process is appropriately validated and suitable corrections are made for attenuation of gamma rays by body tissues. An estimate of regional lung deposition can be made by drawing 'regions of interest' on the scintigraphic image; the precision of this measure is limited by the two-dimensional (2-D) nature of most images which mean that there is an overlay of structures of interest (alveoli, small and large airways), which is most marked centrally. Three-dimensional (3-D) imaging techniques (e.g. single photon emission computed tomography, SPECT, and positron emission tomography, PET) have the potential to give more detailed data on regional lung deposition, but are currently more expensive, employ higher radiation doses, and are less well validated than 2-D (planar) imaging. We consider that, of the available imaging modalities, planar gamma scintigraphy ...