MEASUREMENT OF BACTERIAL SIZE VIA IMAGE-ANALYSIS OF EPIFLUORESCENCE PREPARATIONS - DESCRIPTION OF AN INEXPENSIVE SYSTEM AND SOLUTIONS TO SOME OF THE MOST COMMON PROBLEMS

Computerized image-analysis of epifluorescence preparations is the mos t accurate and simple method for the estimation of bacterial size. We present a simple and inexpensive image-analysis system used to measure and count planktonic bacteria and presently in operation in our labor atory. We show that there is a wide range of image exposures (brightne ss) over which the system performs correctly. Even though the procedur e involves some steps that depend upon operator intervention, the resu lts obtained are highly reproducible and we have estimated the among-o perator variability at 5%. We then discuss the advantages and disadvan tages of different algorithms used for the estimation of volume from t wo-dimensional images and we identify those that perform better for cu rved and unusual cells. We finally estimate that 4 to 6 images and 200 -250 cells are the optimal number of images to be processed and cells to be measured to obtain accurate estimates of population values with the minimum effort. These calibrations should be useful to all those l aboratories that are implementing image-analysis systems.