In a few areas of western North America the principal strain azimuths from geodetic strain measurements differ dramatically from the principal stress directions inferred from fault plane solutions of earthquakes. The reason for this difference may be related to the amount of strain measured by each technique. The strain relieved in an earthquake is about two orders of magnitude greater than that observed by geodetic measurements taken over approximately a seven year time span. Strain measurements may thus sense short-term fluctuations in strain, while earthquakes may average these short-term fluctuations responding to the total strain accumulated to produce a more reliable measure of the longer-term tectonic strain.