Doyle and Campbell Reply: Six terminal measurements are a common method of extracting the in-plane and outof-plane resistivities of high temperature superconducting (HTS) crystals [1][2][3][4]. Usually, the method of Busch [5] is used to extract the resistivities. Levin [1] has developed an approach in which nonlocal effects may be investigated without direct calculation of the resistivities. In the preceding Comment [6], Levin has used this to reanalyze data previously used to infer local behavior in heavy ion irradiated Bi 2 Sr 2 CaCu 2 O 81d crystals [2]. He finds different values for the apparent anisotropy for currents parallel or perpendicular to the ab planes and concludes that there is evidence for nonlocality in all published 6-terminal data on HTS crystals [2][3][4].We have used [1] and the data in [2], together with more accurate estimates of the contact spacing, to evaluate h 2 ͑T ͒ for currents in the two directions from Eqs. (1) and (3) in [6]. The results are presented in Fig. 1. Also shown are the measured values for V 23 and V 67 . The agreement for h 2 above T c , where the voltages are large, is most satisfactory. Just below 75 K, where [6] is focused, h 2 c is indeed larger than h 2 ab as shown by Levin [6]. However, the discrepancy, when a contact spacing of 0.25 mm rather than 0.5 mm is used, is smaller with a ratio of about 2.3 rather than the value of 5.3 in [6]. The discrepancy appears to be systematic and is not easily explained [although the value of h c approaches the limits of validity of the assumptions in Eq. (2)]. Nonlocal behavior is certainly one possible explanation. However, Fig. 1 indicates that the discrepancy in h 2 changes sign just above 75 K where Levin has based his analysis. One would expect that the effects of nonlocality should increase monotonically with decreasing temperature. Furthermore, the discrepancy grows as the voltage levels approach the minimum detectable level in the region between 80 and 88 K. In this region, possible nonlocal effects are expected to be much smaller than at lower temperatures so that the discrepancy is likely to be due to the usual experimental limitations at low signal levels. This also implies that artifactual contributions to the discrepancy cannot be ruled out below 75 K.Two related problems, which have not been previously considered in detail, are the expected direction in which h 2 ͑T ͒ should be shifted in the presence of nonlocal effects, and the expected size of the effect for a given nonlocal correlation length. Levin [6] has pointed out that in all measured cases [2-4,7] the apparent anisotropy for I k c is larger than for I k ab. Although I k c is nominally the force-free configuration, in which flux cutting is difficult since the Lorentz force is small, a significant in-plane component of the current results from the resistive anisotropy since the ab planes are almost