The primary purpose of this study was to determine the causes of extensive divergencies in test results obtained by different operators and different laboratories in the use of the Marshall test for hot mix design and to investigate discrepancies between laboratory-mixed and plant-mixed samples. The data and discussions presented are sufficient to draw some definite and some tentative conclusions. As to the reproducibility of test results, it can definitely be stated that: (1) the experience of the operator affects the results whenever hand mixing is used; (2) the mechanical mixer improves repeatabiliy but shows a tendency to give lower results for stability and specific gravity; (3) heating the breaking heads to 140 F improves test results; and (4) the viscosity of the asphalt at the time of mixing and compaction of the briquet affects the results. Pavement cores and molded specimens of plant mix show similar results. Rutting of a pavement seems to be affected by mixing procedure and by viscosity of bitumen at time of mixing and compaction. Specific gravities of samples mixed at plants are higher than those for mixtures prepared in the laboratory, which affects the apparent optimum asphalt content. Further tests are under way to establish the causes of this discrepancy. The laboratory density obtained by 50 blows of the hammer on each face of the specimen seems to correlate with field density obtained with pneumatic rollers. This study is not complete, and at the present time, definite conclusions should not be drawn. However it is unlikely that a change in the number of blows required for laboratory compaction in design studies will be necessary for present-day traffic. In conclusion, it can be stated that more research is needed to correlate laboratory mixing with pugmill mixing and that repeatability of results in using the Marshall method can be improved by taking several precautionary measures. Considering the conditions encountered in the field and the necessity of quick tests, the Marshall method is a valuable tool in hot mix design and control.