Four of these Particulate Reduction Systems (PMS) were tested on a passenger car and one of them on a HDV. Expectation of the research team was that they would reach at least a PM-reduction of 30% under all realistic operating conditions. The standard German filter test procedure for PMS was performed but moreover, the response to various operating conditions was tested including worst case situations. Besides the legislated CO, NOx and PM exhaust-gas emissions, also the particle count and NO2 were measured. The best filtration efficiency with one PMS was indeed 63%. However, under critical but realistic conditions filtration of 3 of 4 PMS was measured substantially lower than the expected 30 %, depending on operating conditions and prior history, and could even completely fail. Scatter between repeated cycles was very large and results were not reproducible. Even worse, with all 4 PMS deposited soot, stored in these systems during light load operation was intermittently blown-off. Due to these stochastic phenomena the behaviour of these systems is hardly predictable. Furthermore the provision of NO2, through catalysis ahead of the filter or in the filter matrix, is inherent in these systems. Some of this secondary NO2 is emitted. Cost/benefit ratio is high compared to full-flow filters and Diesel engines equipped with partial-flow filters are inferior to SI engines regarding global warming potential. Based on these findings it is concluded that the sustainable performance of partial-flow filters is not yet determined.