To better understand the role of small particles in the microphysical processes and the radiative properties of cirrus, the reliability of the historical in situ measurement database must be understood. A means of establishing this validity is to assume that the in situ measurements are at least consistent, in a broad sense, with the remote sensing data, and vice versa. In this study, an algorithm using Doppler radar moments and Raman lidar extinction is developed to retrieve a bimodal particle size distribution and its uncertainty. Case studies and statistics compiled over an entire year show that the existence of high concentrations in excess of 1 cm^-3 of small particles in cirrus is not consistent with any reasonable interpretation of remote sensing data and is therefore likely from an artifact of the in situ measurement process. This study shows that while the particle concentrations from the Two-Dimensional Cloud Probe generally agree well with the retrieval results, simultaneous concentrations from the Forward Scattering Spectrometer Probe are much higher than the concentrations of small particles implied by the remote sensing measurements. The one-year statistics of the cirrus microphysical properties, including the ice water content, the effective radius and the total particle concentration, show that the occurrence frequency of the concentrations larger than 1 cm ^-3 is below 1%, and, given the possibility of errors in retrieved concentration as large as 100%, this study concludes that the existence of particle concentrations in cirrus in excess of 1 cm^-3 is extraordinarily rare instead of common as suggested by uncritical acceptance of in situ data.