Time series data are commonly obtained by trapping over a standardized period of time, for example daily or weekly. In this paper we present evidence that such sampling designs are inherently irregularly spaced due to the varying developmental rates and population parameters caused by changing temperatures during a sampling season. We modeled an exponentially growing population based on stable fly population growth rates, and then compare different sampling regimes to determine which produces the best estimate of population growth rate. These results are then compared to field data based on weekly sampling at three dairy farms in Ontario over two summers. Transforming catch numbers (N) to ln(N)/(number of degree days within the sampling period) corrects for the irregular spaced sampling in these data. These results support the use of measuring population parameters such as population growth rates in terms of degree days.