_ ROTATING-DISK-SAMPLER MEASUREMENT OF AEROSOLS 227 where k,, the deposition velocity, is defined by the preceding expres- sion and n(r) dr is the ambient concentration of particles in the range r to r-+ dr. The deposition flux may be expressed in terms of the cumulative distribution of particles measured on the electron micro- graphs, ZX(r): dJ =— d Z(r) Ar (5) where A is the area examined on the grid and is the total sampling wo --[S22(ch) time of the rotating disk. Hence, from Eqs. 4b and 5, _ td Zr) 1 where n(r) is the time-averaged distribution defined by Eq. 3. ° The particle diffusion coefficient, D, can be estimated from the Stokes — Einstein expression _kT (7) Dey where k is Boltzmann’s constant and T is the absolute temperature, The coefficient of frictional resistance, f, can be calculated (assuming a Spherical particle) by the expression _ 6rur where pis the viscosity of the medium (air), The semiempirical cor- rection factor recommended by Davies® is C=1+2 (1.257 + 0.400078 9") (9) when the mean free path of the air, A, is given by nM\% A= 0.708v 24) (10) where M is the molecular weight, v is the kinematic viscosity, and R is the gas constant, . Since each sample grid was examined at two magnifications, there are two 2(r) curves for each run, one covering the size range below about 0.02 u and the other the range 0,015 to 0.08 u. In principle, the two sets of X(r) data can be combined to form one Set by the application of a correction factor involving the areas examined at each magnification. This method was rejected since the uncertainties of low particle count around 0.015 u for the smaller particle set and inaccurate sizing around 0.015 » for the larger particle set would magnify the error in the overlap region. For these reasons, it was decided to seek an