The directional sensitivity of unplated slanted‐wire probes is experimentally investigated. It is shown that the yaw response of these probes depends on the roll angle, &bgr;, at which the plane formed by the wire and the prongs is oriented with respect to the yaw plane. At &bgr;=0°, when both planes are aligned with each other, the constant velocity characteristics are monotonic functions of the yaw angle, &psgr;, implying a unique correspondence between the output signals and flow direction. When the two planes are not aligned with each other, a region of yaw angles is found within which the hot‐wire signals become nonuniquely related to the yaw orientation of the probe. The size of this region is shown to depend on the particular value of &bgr;; it increases with increasing &bgr; and vice versa. Problems of nonuniqueness are also found during calibrations of two‐ and three‐wire probes. It is shown that such problems result either from improper orientation of the wires with respect to the calibration plane (in the case of an X probe) or from the arrangement of the wires on the probe (in the case of a three‐wire probe) or both. To avoid nonuniqueness and ambiguous results, the use of these probes is restricted to an experimentally determined range of angular directions of the flow. However, from the current findings it appears that, in the case of a conventional three‐wire probe, the limiting values of the flow angles (angular directions) are relatively small. It is concluded, therefore, that for measurements in three‐dimensional or highly turbulent flows (in which the instantaneous angular directions of the flow are relatively large), an entirely new arrangement of the wires on the probe should be sought.