Instrumental daily maximum and minimum temperatures are reported and archived from various surface thermometers along with different average algorithms in historical and current U.S. surface climate networks. An instrumental bias in daily maximum and minimum temperatures caused by surface temperature sensors due to the different sampling rates, average algorithms, and sensor's time constants was examined using a Gaussian-distributed function of surface air temperature fluctuations in simulation. In this study, the field observations were also included to examine the effects of average algorithms used in reporting daily maximum and minimum temperatures. Compared to the longestrecorded and standard liquid-in-glass maximum and minimum thermometers, some surface climate networks produced a systematic warming (cooling) bias in daily maximum (minimum) temperature observations, thus, resulting biases made the diurnal temperature range (DTR) more biased in extreme climate studies. Our study clarified the ambiguous concepts on daily maximum and minimum temperature observations defined by the World Meteorological Organization (WMO) in terms of sensor's time constants and average lengths and an accurate description of daily maximum and minimum temperatures is recommended to avoid the uncertainties occurred in the observed climatology.