Numerical and experimental techniques were used to characterize aerosol penetration through bends. Agreement between numerical and physical experiments was achieved when the numerical approach was based on the use of a specially developed three-dimensional particle tracking technique. It is also demonstrated that turbulence needs to be included in a particle tracking model. The effect of flow Reynolds number upon particle deposition was examined numerically. Results show that it affects aerosol penetration somewhat; however, it does not appear sufficiently significant to warrant inclusion in any correlation model. For Stokes numbers of 0.07-0.7 and a curvature ratio of 10, the aerosol penetration does not change by more than 5% when the Reynolds number is varied from 3200 to 19 800. Physical experiments were conducted to investigate the effect of curvature ratio on aerosol penetration. The bends were constructed such that each bend had the same initial and final spatial co-ordinates, regardless of the curvature ratio. ANSI N13.1-1969 recommends that the curvature ratio should be at least 10, but the results of this study suggest that the value could be 4. When bends are fabricated from straight tubing, there is a tendency for the tubing to flatten. The effect of flattening on aerosol penetration was tested by pinching bends at the 45°location, with degrees of flattening from 0% to 50%. If the degree of flattening is less than about 25%, it does not have a substantial impact on aerosol penetration. Numerical experiments were carried out to characterize the penetration of aerosols through bends. The geometrical extent of the bends covered only the region of tubing where the radius of curvature is finite. Results were used to generate a correlation model that designate anti users of aerosol transport systems can employ to predict aerosol penetration. The correlation model is valid for the range of Stokes numbers between 0.07 and 1.2, for bend angles from 45°to 180°, and for curvature ratios from 2 to 10.