The varying population density of wet rice arthropods in a 24-h period, at two different locations, was monitored. Using one hill as the sampling unit, 22 categories of arthropods were visually counted every 3 h. In each of two separate studies, at each combination of arthropod category, date and time of sampling, the mean density at each sampling time was calculated and divided by the mean density for that day, to obtain the ratio of population density. Analysis of variance on these ratios (log transformed to normalize residuals and stabilize the variance) indicates sampling time as the only significant (P<0.05) main effect. When combined across categories and dates, ratios for 2100, 2400 and 0300 h were similar, higher than 1, and significantly higher than those from 0600 -1800 h. Examination of each category’s ratio pattern indicates four major behaviour groupings: (1) those with night ratios higher than 1, hence present in higher numbers at night and known to be active then; (2) those with ratios higher at night but known to be inactive then; (3) those -with relatively constant ratios throughout day and night; (4) those with ratios high at certain hours only, usually at night and known to be most active then. Accurate estimation of population should consider the varying density of individuals corresponding with the time of day and changing activity pattern of the arthropod concerned. Hence it is suggested here that the appropriate ratio should be used for the relevant arthropod category when sampling at a certain time of day.