AbstractThe Pacific walrus population has been depleted and subsequently allowed to recover three times in the past 150 yr. As we see it, the population has been made to fluctuate like anr‐selected species, rather than being maintained at a high, stable level, as befits a K‐selected species. The latest depletion began in the 1930s but was not recognized until 25 yr later, by which time the population had been reduced by at least half. Without benefit of communication, the U.S.S.R. and the State of Alaska put similar protective measures into place by 1960, and in the next two decades the walrus population recovered again, at least doubling in size. By 1980, it already was showing density‐dependent signs of having approached or reached the carrying capacity of its environment. As productivity and calf survival declined sharply in the late 1970s and early 1980s the catches more than doubled. We believe that the combined effects of natural curtailment and human intervention may be bringing the population down again rather rapidly. With the present, crude monitoring methods, delayed management responses, and poor international communications, however, the downward trend may not be acknowledged for at least another decade, by which time the unilateral Soviet and American corrective measures are likely to be too much, too late. Walrus management needs to be based less on response to immediate crisis and more on long term prediction than it has been in the past. Because the U.S.A. and U.S.S.R. are trying to manage the same walrus population, without sufficient communication or consensus and sometimes to opposite ends, an international joint management program needs to be implem