The results from pulse testing were used to predict waterflood performance in the Kelsey field. This prediction was verified by field waterflood performance. The tests showed the presence of an unmapped sealing performance. The tests showed the presence of an unmapped sealing discontinuity and the fact that all but one major fault were nonsealing. Introduction The pulse-testing technique described by Johnson et al. and later used in an extensive application in an oil-producing field has proven to be a useful tool for describing areal reservoir heterogeneities. This paper describes the use of pulse testing to predict the performance of a planned waterflood in Exxon Co., U.S.A.'s Kelsey field in South Texas, a producing oil field containing a number of mapped faults. The wells are completed in the Frio formation, which is a nonmarine, river point-bar, shaly-sand type formation. This flexuretype field is produced on the downthrown side, and the original production of the 38- to 42- degrees API gravity oil was by gas-cap expansion. Even though the main purpose of pulse testing was to determine reservoir connectivity, pulse testing was to determine reservoir connectivity, values for storage, transmissibility, and hydraulic diffusivity were obtained for some of the between-well reservoir properties. This paper also summarizes the history of a subsequent waterflood, performance of which verifies the pulse-test results, Pulse-test values of transmissibility, storage, and hydraulic diffusivity reported in this paper were within the range of single-well test, core, and fluid property data estimates of formation properties of this South Texas field. Results of feasibility studies, based on these data, helped in the design of the pulse-testing program described here. pulse-testing program described here. Pulse-Test Procedure and Pulse-Test Procedure and Evaluation of Results The pulse test requires two wells - a pulsing well and a responding well. The scheme is to create a sequence of rate changes in the flow at the pulsing well and measure the resulting pressure changes at an adjacent responding well with a very sensitive wellhead pressure gauge. The measuring sensitivity of the gauge used for these tests was 3 x 10 to the -4 to 12 x 10 to the -4 psi. The precision of the gauge is dependent on the surface wellhead pressure of the responding well, which, in most cases, was about 300 psi. For our tests, the pulsing wells were flowing, producing oil wells with rates of 175 to 880 RB/D of oil. Two exceptions, Wells 11 and 16, shown in Fig. 1, already had been converted to water injection. Therefore, we used them as pulsers with injection rates of 1,400 and 2,000 BWPD, respectively. The tubing in the responding wells was loaded with lease crude to provide a gas-free liquid with a positive pressure to the surface. Then the sensitive pressure gauge was connected to the well tubing and was allowed to equilibrate for 12 hours before pulsing was begun. A pulse was started and continued until a definite response could be observed, or for period of 12 hours or longer but not more than 1 day if period of 12 hours or longer but not more than 1 day if there seemed to be no response. Table 1 gives the well pairs tested, distance between them, pulse rate, pulse pairs tested, distance between them, pulse rate, pulse interval, and gauge sensitivity for each test. In general, these values are the same for both the response and no-response cases. However, the pulse intervals were longer in the no-response cases in an attempt to enhance the chances of detecting a response. An example of the data where a response was obtained between a well pair is shown in Fig. 2. Fig. 3 shows an example where no response was observed. JPT P. 914