Trenching survey across a fault is the best way to obtain informations on the timings of past earthquakes. Profound implications for seismic-hazard evaluation and active-tectonic processes call for numerous trenching surveys. In this paper, some problems on event analysis in current trenching surveys are discussed. This is indispensable for making standards for trench investigation. The main conclusions are the following;
1. Structures indicative of activity of a fault should carefully be distinguished from those related to liquefaction. Heavy shakings caused by other faults close to the fault under investigation might have produced sand-blows, injected sand dikes, intense rumplings and so on. It is particularly required in the case of argument of fault segmentation.
2. Talus or colluvial apron derived from a fault scarp, and a drastical facies change in subsequent deposits do not always indicate a faulting event.
3. Short branch faults are not good indicator bracketing of the time of an earthquake.
4. When recognizing plural events, particularly based on the grade of deformed strata, one should pay much attentions on the accumulation of displacements.
5. Humus and peaty materials can be piled up on a slope. Therefore, it is necessary to observe carefully whether these strata are deformed or not.
