Continuous seismic recording is fundamental in shaping our understanding of earthquake kinematics, dynamics, and hazards. However, most modern computational analysis methods can only be applied to digital records which only became available in the late 1970s. The decades since represent far less time than the typical recurrence intervals between large earthquakes on most faults, thus drastically undersampling the seismic cycle. On the other hand, analog data have been recorded on paper for over a century, but they were conventionally used to determine event timing, origin, and magnitude only, leaving the focal mechanism and kinematic aspects of many historical earthquakes unknown or uncertain. Therefore, the vast collection of analog seismograms remains an under-explored treasure. Revisiting these legacy data is an endeavor which requires great effort but reaps great rewards. Yet, most scholars who might be interested in studying past events are unaware of the existence of these records or find the process of acquiring and digitising historical seismograms prohibitive. Due to the lack of usage, many archives lack necessary support such that the vast amount of legacy data could soon be lost without intervention. To promote the preservation and analysis of this precious dataset, this chapter gives a state-of-the-art overview of methods of earthquake source parameter determination using analog seismic records, with the aim to help interested readers embark on projects making use of historical seismology.