a pulse fishing scheme, applied over a three year period (1980 to 1982), deliberately removed 3226 walleye from henderson lake, ontario, causing the stock to collapse. the objective of the removal was not only to seek out characteristics which could have served as a predictor of the collapse, but also to test the applicability of pulse fishing as a management alternative for walleye. several indices failed to serve as indicators of the population collapse. at high densities of walleye the schumacher-eschmeyer population estimate (using trap nets) apparently only estimated the brood stock abundance. not only did it incorrectly indicate the walleye population to be relatively stable after two years of intensive removal, because of younger fish being recruited to the gear, it also failed to forewarn of the impending collapse of the stock in the third year. catch per unit effort data proved a poor index of fish abundance for northern pike and white sucker. walleye abundance was significantly correlated to walleye catches from four and six-foot trap nets. catches in both eight-foot trap and gill nets were very poorly correlated to fish densities, although these gears were used at reduced effort levels. evidently vulnerability to gill nets of the remaining walleye may change as a result of reductions in food abundance causing more predator movement to seek prey. changes in fish condition with changes in population abundance could also influence gill net capture rates. condition of most walleye age-classes and both walleye and northern pike fecundity showed a time lag in response to fishing, not increasing significantly until 1984, 3 years after stock collapse. walleye growth rates increased only for younger age classes (ii - v). both mean age and the mean age to maturity decreased with exploitation. annual production estimates appeared to serve as a good indicator of walleye response to fishing. they originally ranged between 1.92 to 3.07 kg/ha/yr before exploitation, quickly fell to negative values following the second year of exploitation, and only increased to 0.41 kg/ha/yr by 1984, two years after exploitation ceased. the most promising predictors of the walleye population collapse were length at age increases, abrosov's mean age to mean age at maturity index (0.5 critical value), annual production estimates and possibly petersen population estimates. northern pike and white sucker numbers have not increased since walleye exploitation began. they appeared not to be filling any vacated walleye niche, at least over the short term. due to an unfortunate succession of abiotic factors unfavourable to production of strong year-classes, walleye recruitment was very poor. this demonstrates that any implementation of an episodic removal scheme must take into account such possibilities. further research in this area should clarify the exact sequence of removal and recovery necessary to successfully implement this scheme for the management of walleye.