The ability of foam to divert gas flow over a long period of gas injection in a Surfactant Alternating Gas (SAG) foam process is important for the economics of foam-diversion processes for enhanced oil recovery. Here we interpret field data from the foam test in the Cusiana field in Colombia, South America (Ocampo et al., 2013). In this test surfactant was injected into a single layer that had been taking about half the injected gas before the test; then gas injection resumed into all layers. Based on the size of the surfactant slug injected and estimates of adsorption and of water saturation in the foam in situ, we estimate that the treated region extended about 5.3 m from the injection well: fortunately the results to follow are not sensitive to this estimate. Based on the change in injection logs before the test and at day 5 of the test, when approximately 30 pore volumes of gas has been injected, foam still reduced gas mobility in the treated layer by about a factor of 9. We base this estimate on the decrease of injection into the treated layer and the increase into the other layers; the results are consistent among the layers. After 35 and 152 days of injection (220 and 1250 pore volumes gas injected), foam reduced gas mobility in the treated zone by about a factor of 4 and 2, respectively. This result suggests that foam continued to reduce mobility by a modest amount even after long injection of gas. In this test, the large volume of gas had quickly penetrated far beyond the edge of the surfactant bank. In a design where a larger bank of surfactant were injected, a much greater and longer diversion of gas would be expected. On the other hand, foam did weaken progressively as it dried out. Foam models where foam remains strong at irreducible water saturation would greatly overestimate foam effectiveness at long times in this test.