The material presented in this thesis falls into three main sections: 1) The measurement of nebulizer suction and its applications. Three methods of measurement of nebulizer suction were evaluated. The potential and limitations of each method was assessed. The most reliable and generally applicable method was measurement with a mercury manometer via a T-piece during aspiration, but after correction for suction drops along all the nebulizer capillaries and across connections. The measurement of suction was then applied to practical problems in AAS, i.e. the lack of dependence of signal on sample solution temperature, and to provide immediate warning of drift in aspiration rate. 2) Observations and causes of deposition in spray chambers. The deposition patterns of aerosol lost in the spray chamber was studied using a lithium tracer. Areas of high turbulence were identified and their positions related to the deposition patterns. The nebulizer/spray chamber system was considered to fulfill a sub-sampling role and the dual roles of pneumatic nebulizer as pumps and sub-samplers were critically discussed. 3) Observations on impactors in flame AAS. Five techniques for the evaluation of the effects of impact beads and other impactors were evaluated, i) aerosol droplet size distributions ii) aspiration of dye solutions iii) aerosol sizing with a second species introduced through the bead iv) effects on linear absorbance range v) effects on the extent of chemical interference The advantages and disadvantages of each technique were considered and the value of impactors critically appraised. An assessment was then made of the extent to which an impact cup or bead may be used to regulate sensitivity in flame AAS with a view to increasing the useful working ranges of calibration graphs.