The impact of the inlet conditions on particle dispersion in a post-transitional shear layer is investigated using two-dimensional simulations. The flow inlet conditions are varied between a monotonically changing inlet velocity profile (error-function) and one that exhibits a wake deficit. The particle field inlet conditions are varied by altering the particle inlet location. Monodisperse, dilute particle fields are considered. The numerical models for both the flow and particle fields are unsteady and Lagrangian. The applicability of a simplified form of the equation of particle motion is verified by comparing its solutions with those of a more comprehensive form. The maximization of dispersion for particles with intermediate Stokes numbers documented in previous studies is reproduced. So is the fact that increased proximity of particle introduction to the shear flow leads to enhanced dispersion. For the error-function flow, particles introduced close enough to the shear flow so that they get entrained into the vorticity region, disperse more if they originate from the fast stream. Particles introduced further away from the shear flow, on the other hand, disperse more if they originate from the slow stream. These results persist for all Stokes numbers and are related to the flow dynamics. They also help reconcile the contradictory results of previous studies. The impact of introducing the particles from either of the two streams is much more significant when a wake-modified inlet velocity profile is used. Dispersion patterns show a strong asymmetry in favor of the fast stream. This is attributed to the presence of positive vorticity on the slow speed side of the shear layer. This vorticity has its origin in the upstream boundary layer and tends to weaken the local particle entrainment. Overall, the results of this study suggest that the optimal location of particle introduction is on the fast stream side and close to the splitter plate. These results also imply that simulations that ignore the wake component of the inlet velocity profile may yield misleading results.