Field measurements in the past several years have documented electromagnetic signals which are attributed to precursory stress and strain changes which were followed ultimately by earthquakes. Precursory electric field changes observed in Greece on multiple dipoles have been used to issue earthquake predictions. While the source of these signals is still unknown, a sufficient number of predictions has been issued to allow some, but not all, statistical analyses to show this method is better than randomly sampling the earthquake catalog. Ongoing efforts to identify the sources of both these signals and the magnetic field variations prior to the Loma Prieta earthquake are focusing on electrokinetic coupling of fluid flow and transient electric fields. A mechanism related to local fluid flow appears to be best suited at this time of explaining the variety of purported precursors. However, much more work is needed to improve the observations and refine the models of precursor generation. Efforts to monitor magnetotelluric transfer functions at longer periods ( T>10s) have been hampered by variability of the functions. The use of modern noise reduction techniques such as remote referencing should reduce this variability, but may not reduce errors to a level needed for monitoring. Monitoring of high frequency (81 kHz) seismoelectric emissions may be promising, but lack of simultaneous observations on multiple instruments hinders the utility of this technique.