Random distributions for a wide range (1–100,000) of chironomid head capsules (HC) were simulated on a 1-m2 surface. The number of HC found in circular surfaces equivalent to standard core diameters (90 and 63 mm) was estimated 1000 times, over the range of tested densities. For each number of HC found in the samples, the range of simulated densities was estimated using a threshold probability (p > 0.95). This enabled us to develop equations to infer HC density from sample counts. Because of the threshold probability for comparable sample counts, the equations yield higher estimated densities under a random distribution than for a regular distribution. The probability of sampling at least one HC was >0.95 for densities of 900 HC m−2 for the 90-mm core and 1400 HC m−2 for the 63-mm core. For a specific sample count, the range of actual densities was ~10 times higher for the 63-mm core than the 90-mm core. Comparison with field larval densities revealed that most densities were too low to be suitable for annually resolved reconstruction of a quantitative signal, using current corer sizes, although a large number of populations can support sub-decadal analyses. Nonetheless, some lakes exhibit population sizes large enough to reconstruct robust quantitative estimates of past chironomid abundances. This work provides guidance to reconstruct species dynamics and fine-scale time series analyses in paleoecology.