Limitations of the stochastic approach, a method widely used in simula
ting droplet dispersion in turbulence, are examined using two well-def
ined two-phase, turbulent flow problems. It is found that the stochast
ic approach is restricted to flows in which the continuous and dispers
ed phases can meet each other's correlative condition. Also, use of th
e steady-state formula of the drag coefficient in the Lagrangian equat
ion of motion for models based on the stochastic approach could cause
significant errors in the prediction of complete two-phase turbulent s
tructures, particularly for large droplets.