Differentiating intracellular from extracellular enzyme activity is important in soil enzymology, but not easy. Here, we report on an adjusted sonication method for the separation of intracellular from extracellular phosphatase activity in soil. Under optimal sonication conditions [soil:water ratio = 1/8 (w/v) and power density = 15 watt ml(-1)], the activity of alkaline phosphomonoesterase (phosphatase) in a Haplic Cambisol soil increased with sonication time in two distinct steps. A first plateau of enzyme activity was reached between 60 and 100 s, and a second higher plateau after 300 s. We also found that sonication for 100 s under optimal conditions activated most (about 80%) of the alkaline phosphatase that was added to an autoclaved soil, while total bacteria number was not affected. Sonication for 300 s reduced the total bacteria number by three orders of magnitude but had no further effects on enzyme activity. Our results indicate that the first plateau of alkaline phosphatase activity was derived from extracellular enzymes attached to soil particles, and the second plateau to the combination of extracellular and intracellular enzymes after cell lysis. We conclude that our adjusted sonication method may be an alternative to the currently used physiological and chloroform-fumigation methods for differentiating intracellular from extracellular phosphatase activity in soil. Further testing is needed to find out whether this holds for other soil types.