A program was undertaken to develop a statistically based sampling scheme for filter manufacturing. The basic concepts of statistical process control were adapted specifically to the filter rodmaking process. Recommendations are made for test sample size, sampling frequency, and process control limits. In addition, techniques are reviewed for interpretation of the resulting control charts. Use of the recommended sampling scheme and control charts will maximize production of filter rods within specification limits and minimize process over-control. The process capability, i.e., the amount of natural variation in the process, was determined for the production of filter rods from cellulose acetate tow. A range of tow items were characterized and the process capability was determined for each of several filter parameters. The parameters include encapsulated pressure drop, weight, circumference, and filter firmness. The process capabilities and selected specification limits were compared to determine whether the process would be capable of meeting the specifications. In addition, taking both the process capability and assumed specification limits into account, recommended sample sizes and sampling frequencies were determined. In determining sampling size and frequency, an"acceptable"risk of failing to identify that a process is out of control is assumed. Finally, experiments were run to perturb the rodmaking process and operate in an"out of control"mode. Using the recommended sample size and frequency, control charts were generated to determine whether the process would be recognized as being out of control.