Astrometry with NPOI is based on the measurement of fringe delays, which are related to the baseline orientation and star position. There are three different corrections to the uncalibrated FDL delays, which are due to atmospheric refractive index fluctuations, path length variations in the feed system, and motions of the siderostat pivots. We call the application of these corrections to the raw FDL delays calibration of the delays. $\cal OYST\!ER\,$, similar to the visibility calibration, has storage for three calibrated delays, i.e. group, dry, and wet delays. These are the result of the dispersion correction procedure (dispcorr), which computes cumulative corrections based on the white light fringe position (group delay), the dry and wet air dispersion constants, respectively. All subsequent metrology corrections are only applied to these three calibrated delays. Note that $\cal OYST\!ER\,$does not have separate variables for delays with and without metrology corrections, for example. The user has to keep track of the actual calibration status of the delays.