The unique conditions for the application of laser-induced breakdown spectroscopy (LIBS) as a metal emissions monitoring technology have been discussed. Because of the discrete, particulate nature of effluent metals, the utilization of LIBS is considered in part as a statistical sampling problem involving the finite laser-induced plasma volume, as well as the concentration and size distribution of the target metal species. Particle sampling rates are evaluated and Monte Carlo simulations are presented for relevant LIBS parameters and wastestream conditions. For low metal effluent levels and submicrometer-sized particles, a LIBS-based technique may become sample limited. An approach based on random LIBS sampling and the conditional analysis of the resulting data is proposed as a means to enhance the LIBS sensitivity in actual wastestreams. Monte Carlo simulations and experimental results from a pyrolytic waste processing facility are presented, which demonstrate that a significant enhancement of LIBS performance, greater than an order of magnitude, may be realized by taking advantage of the discrete particulate nature of metals.