Counting Bacterial Colonies

MALLIGO, JOHN E. (Fort Detrick, Frederick, Md.). Evaluation of an automatic electronic device for counting bacterial colonies. Appl. Microbiol. 13:931-934. 1965.-An automatic colony counter was tested extensively with colonies of two bacterial species, Serratia marcescens and Bacillus subtilis var. niger, grown on agar media. A stable relationship was established between machine counts and counts done visually by technicians. The calibration curve and estimates of the efficiency of the machine are presented and discussed. It is estimated that a 40% reduction in colony counting time is feasible through-use of the machine. In recent years, advances in electronic techniques have resulted in the development of many automatic and semiautomatic devices covering many aspects of the biological and physical sciences. A notable factor in these trends to automation has been the development of highspeed scanning systems. Some scanners, such as those of Roberts and Young (1952), Tolles and Bostrom (1956), and Sawyer and Bostrom (1958), have been specially adapted for use in the biological sciences. A noteworthy application is the automatic counter of microbial colonies developed as a cooperative effort between Fort Detrick and the Allen B. Du Mont Laboratories. This counter is shown in Fig. 1, and the technology was described by Mansberg (1957), Mansberg, Yamagami, and Berkley, (1957), and Alexander and Glick (1958). A practical, as well as interesting, aspect of the development of the automatic colony counter was the statistical evaluation and the resulting correction equations described in this report. MATERIALS AND METHODS The bacterial cultures used were Bacillus subtilis var. niger grown in casein acid digest medium (Borden Co., New York, N.Y.) and autolyzed to eliminate vegetative cells, and Serratia marcescens suspensions grown in tryptose-phosphate-cerelose medium. S. marcescens and B. subtilis were plated on peptone agar (Wilson & Co., Chicago, Ill.) and Tryptose Agar (Difco), respectively, to provide large numbers of plates with varying numbers of colonies. The range from 30 to 300 colonies per plate was covered in 30-colony increments to provide a realistic distribution of sample plates to be counted. Care was taken in spreading the inoculum to avoid the growth of colonies on peripheral areas of the plate. The counter was adapted to count colonies 0.5 mm in diameter and larger. Although the lower size threshold of colonies which the counter will detect may be varied to sense objects as small as 0.1 mm in diameter, errors are introduced by the system becoming sensitive to extraneous targets, such as dust and lint or minute scratches in the agar or petri dish. The problems associated with lowering the size threshold of the device were discussed by Mansberg (1957) in some detail. RESULTS AND DISCUSSION Early experimentation with the automatic colony counter showed that the machine was able to count, with great accuracy and precision, test plates with plastic spots simulating bacterial colonies. When plates with living cultures were used, however, the machine count was lower than the visual count when more than 100 colonies were present on the plate. This difference in count between the technician and the machine was directly proportional to the number of colonies on the plate. An intensive investigation was conducted to define accurately the observed bias and to prepare a calibration curve of general applicability. Table 1 shows the relationship between the machine count and corresponding visual counts. At plate count levels of 300 colonies per plate, the automatic counter estimated about 83% of the colonies observed by the technician. Preliminary study showed that there were four causes for this bias: (i) colonies on peripheral area of the agar, (ii) optical imperfections and scratches in the agar or petri dishes, (iii) machine 931 on N ovem er 2, 2017 by gest ht://aem .sm .rg/ D ow nladed fom