Membrane filter samples have been loaded with controlled amounts of fibers, and the samples prepared for counting by phase contrast optical microscopy. Samples have been counted by an automated image analyzer and by experienced fiber counters in three laboratories. The counts show a bias in the estimation of fiber concentration which is related to the number of fibers per unit surface area of the filter sample. Generally, the form of the bias is toward underestimates at very high densities and overestimates at low densitites. The overestimates at low densities are attributed to some extent to counts of background features of the filter as fibers. However, there were also indications that if counters do not record background counts then they are likely to produce underestimates at low densities. The underestimates at high densities are attributed to the difficulties caused by the large number of fibers in each field. The recommendation from these findings is that sample volumes should be selected so as to produce sample densities in the range 100 to 1000 fibers/mm2 in order to reduce bias in fiber counts.
No takes yet. Share an insight, caveat, or question.
Cherrie et al. (1986) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: