Evaluation of the COLI-BART system, a novel method for the quantitative detection of total and fecal coliform bacteria in water.

Evaluation of the COLI-BART system, a novel method for the quantitative detection of total and fecal coliform bacteria in water. Abstract Roy Cullimore and Lori Johnston. Droycon Bioconcepts Inc, Regina, Saskatchewan Detection of the coliform bacteria in water remains a very critical method for the determination of health risk. In the last two decades there has developed a need to be able to detect the total or fecal coliform bacteria at the presence / absence (P/A) level in a manner that could be applied at remote locations as well as in the microbiology laboratory. A number of test systems now exist that can undertake this at the P/A or semiquantitative level but quantitative tests still rely on either a version of the membrane filtration (MF) or the most probable number (MPN) technique. The system developed here is the patented COLI-BART system that utilizes the principles of the MPN method (gas production by fermentation) but detects the population size by the rate at which the gas is formed in the test rather than the number of tests in a dilution range that produce gas. Here the premise is that the faster the gas is produced, the greater the original population of coliform bacteria in the water sample. Experiments were conducted using ATCC pure cultures of Escherichia coli and other coliform bacteria as well as using natural samples with known coliform populations. It was found that the time lag measured in seconds by the elevation of a patented gas entrapment thimble could be related to the population size with an accuracy of R 2 of better than 0.8 in a logarithmic regression analysis. With natural samples it was demonstrated that the time lag could be used to generate the most probable population (MPP) with statistical accuracies similar to that of the MPN method. This test method has been found to be functional in the field as well as the laboratory and confirmatory studies are achievable where the completed positive test is returned to a suitable accredited laboratory. Historical Perspective It was as early as 1658, when Athanasius Kircher, a Jesuit priest, first documented a microscopic analysis of the nature of putrifaction. This investigation led to the discovery of bacteria, tiny organisms that can not be seen with the naked eye. Bacteria were first detailed by such great names as Leeuwenhoek, Koch, Pasteur and Lister. It was determined that through sanitation, many of these organisms could be controlled under certain circumstances. Groups of organisms were determined through their function or causation of disease. (Carter and Smith, 1953) The testing for bacteria commonly referred to as coliforms, has been going on for over 100 years. In 1885, colon bacteria, or coliform bacteria in today s common rhetoric, was not a single species of bacterium, but a name applied to a group of organisms whose appearance is too similar to permit differentiation. Many of these bacteria were 1

characterized by the available media in which they were cultured. In media containing peptone, idol is formed, where sugar was found, large volumes of gas was produced, sufficient to break apart the media. This can be confirmed through the use of a fermentation tube. Milk and Litmus solutions were also used in the early cultures, producing coagulation and where large quantities of acid was produced, discoloration occurred. (Zapffe, 285) By 1910, further evidence of the colon group, also called lactose fermenters, was confirmed. During this period in history, laboratory techniques remained similar with the use of milk and carbohydrate, particularily sugar, media. Acids produced include lactic, acetic, formic and succinic acids. The fermentation of sugars produce acid and gas, which was a definitive characterization of this group. Nearly all bacteria, 80%, within this group produce gas from the fermentation of lactose. (Park, 255-260) It was also noted that the process of fermentation is not simply a hydrolytic process, but one in which the carbon and oxygen are sundered and formed. The change is through a process of breaking down the sugar molecules, a true decomposition. By the mid 20 th century, the colon group, or coliform group of bacteria continue to be a major issue. New techniques were developed to test for the presence of these organisms. The tests are based on the fact that the coliform organisms belong to a group which ferments lactose with the formation of acid and gas. The testing procedure is broken down into three steps, the presumptive, the partially confirmed and the confirmed test. The presumptive test examines samples for gas formation. If gas is formed it is considered a positive reaction and interpreted as containing lactose-fermenting organisms which may be coliform bacteria. The partially confirmed test consists of streaking the broth culture from step one onto a special lactose agar, such as eosin-methylene blue agar. Differential growth of the organisms assists in the separation of colonies. To confirm the presence of coliform bacteria, the specific colonies grown on agar are transferred to either agar slants or broth and Gram stain and microscopic analysis is done. (Thompson, 1954) There are many disparities when examining, in a quantitative manner, the coliform group. Traditional techniques such as the agar slants or plates, have inherent problems. Historically, the plating techniques have allowed for isolating organisms, obtaining pure cultures and classification or identifying individual types of bacteria as well as enumerating them. This technique is based on the assumption that when a microbial cell grows on solid medium, and care is taken to avoid cell aggregation and excessive cell numbers, that each individual cell will grow and produce an isolated colony Detection of the coliform bacteria in water remains a very critical method for the determination of health risk. In the last two decades there has developed a need to be able to detect the total or fecal coliform bacteria at the presence / absence (P/A) level in a manner that could be applied at remote locations as well as in the microbiology laboratory. A number of test systems now exist that can undertake this at the P/A or semiquantitative level but quantitative tests still rely on either a version of the membrane filtration (MF) or the most probable number (MPN) technique. The system developed 2

here is the patented COLI-BART system that utilizes the principles of the MPN method (gas production by fermentation) but detects the population size by the rate at which the gas is formed in the test rather than the number of tests in a dilution range that produce gas. Here the premise is that the faster the gas is produced, the greater the original population of coliform bacteria in the water sample. Experiments were conducted using ATCC pure cultures of Escherichia coli and other coliform bacteria as well as using natural samples with known coliform populations. It was found that the time lag measured in seconds by the elevation of a patented gas entrapment thimble could be related to the population size with an accuracy of R 2 of better than 0.8 in a logarithmic regression analysis. With natural samples it was demonstrated that the time lag could be used to generate the most probable population (MPP) with statistical accuracies similar to that of the MPN method. This test method has been found to be functional in the field as well as the laboratory and confirmatory studies are achievable where the completed positive test is returned to a suitable accredited laboratory. Introduction Coliform bacteria form a part of the enteric bacteria. Particularly important from the health risk viewpoint is that other members of the enteric bacteria play a significant role through causing diseases such as typhoid, food infections (salmonellosis), and bacterial dysentery (shigellosis). In addition there are many other genera of significance as plant pathogens, the causal agents of food spoilage, and of contributing to major activities in the environment. While they can be split into five tribes (Table One), the enteric bacteria are all gram negative rods that are catalase positive and, where motile, possess peritrichous flagella. Most genera will reduce nitrate to nitrite (denitrification). Some of this family are simple chemoorganotrophs able to use glucose as the sole source of carbon and energy, while others are very fastidious requiring various vitamins and amino-acids. All can grow in the presence or absence of oxygen. Table One Differentiation of the five tribes of the Enterobacteriaceae Tribe # I II III IV V Habitat Primarily parasitic in animals Primarily saprophytic, some can be pathogenic Saprophytic Plant pathogens Pathogenic using vectors for transmission The fecal coliform bacteria are represented in Tribe I, consists of five genera, all of which are primarily parasitic in animals with some causing serious diseases particularly in humans. The genera include: Escherichia, Edwardsiella, Shigella, Salmonella and Citrobacter. Escherichia includes many relatively benign strains that parasitize the 3

intestine (gastro-enteric tract) of many animals and are particularly dominant in the human species. Some are opportunistic pathogens and can cause forms of diarrhea. Edwardsiella is a common parasitic genus found in reptiles including snakes, toads and frogs. There are many similarities between Escherichia and Edwardsiella that makes differentiation challenging. Shigella is an intestinal pathogen causing bacillary dysentery. Salmonella causes three major infections related to enteric fevers (typhoid, paratyphoid), food infections that can cause gastro-enteritis, and also septicaemia. Citrobacter is essentially a border line genus with the tribe II and includes a range of opportunistic pathogens and also some commonly found in the environment. Fecal coliform bacteria are associated primarily with Escherichia coli in tribe I but some tribe II genera can also trigger as fecal coliform bacteria during coliform testing.. It is interesting to note that the sugar lactose is not commonly found throughout the environment but appears to be synthesized the most commonly in mammalian milk. Mammals when suckling during infancy therefore take in large quantities of lactose in the milk which biases the microflora in the gastroenteric tract to those bacteria that can anaerobically ferment lactose (i.e., coliform bacteria). These coliform bacteria also have to be able to resist the inhibitory effects of bile salts (excreted into the gastroenteric tract) and be able to grow at warm blooded temperatures (e.g., 35 o C). Because of these restrictors, coliform bacteria may be selectively grown in culture conditions where other contaminant bacteria would be suppressed. The restricting factors are the use of lactose as the major energy source; application of bile salts to selectively restrict competition to those bacteria normally found growing in the gastroenteric tract; and the use of an above environmental norm temperature for growing these coliform bacteria (i.e., 37 o C). Acid products are easily determined by colour shifts in ph indicators and the gas may be entrapped for direct or indirect observation 10. Where coliform bacteria are detected then confirmatory tests need to be performed if the presence of Escherichia coli is to be determined. The need to confirm that E. coli is present is important because some of the other coliform group are able to not only survive but also grow within the natural environment. Such microorganisms in a water system would therefore cause a positive coliform test. That test however remains presumptive until the presence of E. coli is confirmed. The most common genera causing these types of interferences are Enterobacter and Klebsiella. Massive increases in the human population and greater exploitation of fresh waters have led to an increasing public sensitivity to the hygienic quality of potable water. Even though many alternative methods for defining hygiene risk in water have been proposed, coliform bacteria have stood the test of time as a reliable indicator group. Consequently the industry to detect coliform bacteria has grown substantially and has been subjected to the latest molecular / immunological test methods such as the ONPG (coliform bacteria) and MUG (E. coli) tests. The recent events at Walkerton, Ontario and North Battleford, Saskatchewan delivered a message that a simple, reliable and robust coliform test was still needed particularly if it could have the following characteristics: Used in remote locations Could be performed by persons without extensive training Could provide positive tests that could then be easily subjected to traditional confirmation using the standard laboratory techniques. 4

Methodology Easy to interpret at the visual hands-on level but could also be easily integrated into a computer-controlled package where the data could be generated locally and observed remotely with a provable chain of custody. Long shelf life under minimal storage conditions. To be recognized, the COLI-BART system needs to deliver data comparable to the standard MPN test method with at least the same level of sensitivity to the detection of the targeted coliform bacteria group. This would require a successful statistical comparison of the time lag (TL generated in seconds to fermentation gas generated thimble elevation) to the statistically calculated population generated by the standard MPN method. The following objectives therefore form a part of the objectives to determine the validity of the COLI-BART system: A. The test would be fully quantitative and will generate a most probable population (MPP as opposed to an MPN) estimate for the population size of the coliform bacteria in the sample. B. Can be conducted at remote locations (with an assured repeatability and satisfactory chain of custody). C. Operator time is reduced to less than five minutes for initiating the test procedure. D. Test completed in 24 hours for the total coliform and 30 hours for the fecal coliform. E. Involves using a single test utilizing 100ml of water sample to conduct a quantitative test. F. Population of coliform bacteria is predicted on-screen as soon as the activities of the coliform bacteria are detected (usually from 4 to 27 hours depending upon the population size of coliform bacteria being detected in the sample). G. The COLI-BART readers would have a one-year over-the-counter replacement warranty and would be expected to last a minimum of three operating years. See product specifications (3.6) for more information on the readers. H. While the reader has been designed for use under common conditions where conditions are dry and cool, the reader has not been protected against extremely wet and corrosive conditions. Due diligence should be employed in the location of the reader. I. 100ml of sample is added to initiate the test and this act dilutes the selective medium to the correct strength for the particular coliform bacteria test being conducted. J. All completed COLI-BART testers would be returned to a laboratory for disposal (even where negative for coliform bacteria) with the option, where positive, for subsequent confirmation of the presence of the coliform bacteria in the culture created by the test. K. The double-wrapping of the completed COLI-BART tester provides additional protection from accidental spillage. This would be an advantage to the user since this would also provide a safe method of disposal or laboratory confirmation of the coliform bacteria present in the sample if required. L. Storage life for a COLI-BART tester is three years when stored in a sealed aluminum foil pouch (standard practice in all BART testers). It is recommended that the product 5

be stored in a cool dry environment. M. Variability resulting from minor differences between the various operators performing the coliform tests should be reduced since the number of operatorinitiated steps are reduced to simply filling the tester, burping the thimble, placing the tester into an available channel in a reader and initiating the test. Disposal parallels methods already in common use. N. To reduce risk to the human handling the container, the COLI-BART tester is provided with a second flexible plastic bag that is sealed once the sample has been added. This outer bag acts as a secondary barrier to prevent the release of any cultured material being generated within the test. Contact time for the operator setting up the test is minimized. O. A chain of custody is assured through a unique serial number that is etched onto the tester. Each tester would be shipped within an outer container (when not use in a laboratory) that would allow all used vials to be returned to a designated laboratory for any confirmatory work that may need to be conducted and safe disposal. The only direct exposure to the contents of the tester would be at the point of filling prior to sealing the outer protective bag and insertion into the reader. Shipping of new testers would be through normal courier services with the testers already in the containers that would be used to ship the completed test back to the laboratory for disposal. The chain of custody would be from the manufacturer to the wholesaler to the end user who returns the completed tester back to a suitable disposal / confirmatory facility. The serial number for each tester would allow the tracking of each tester through the chain of custody. P. All operators and operators of the COLI-BART system would have to go through a training program to assure that the testing process was conducted in a suitable manner. The principles of the COLI-BART system are based upon the ability of coliform bacteria to ferment selective culture media with the production of gas. Traditionally the gas, where it is used as the indicator of coliform presence, is collected within an inverted glass (Durham's) tube. The gas collects as a visible bubble at the upper (closed) end of this tube. Population assessment is made through a statistically appropriate multiplicity of diluents of the water sample in which only some generate gas. This method is known as the multiple tube method and generates a most probable number (MPN) statistically as an indicator of the size of the coliform population. In this technique, the Durham's tube is sometimes difficult to view under test conditions due to clouding of the culture medium. In the COLI-BART tester, the Durham's tube is replaced with a patented gas D thimble. The unique feature of this device is that it consists of a series of slots within an inverted plastic container having a density that is greater than the culture medium with the water sample added. This allows the D thimble to sink. The (patent pending) version II of the device allows for a free movement of cells and chemicals through the slots between the D thimble device (version II) and the rest of the liquid medium. This is not possible in the glass walled Durham's tube. When gases are produced by fermentation, some of theses gases will collect inside and reduce the density of the device. As the gas forms, therefore, the D thimble rises up (as the density drops) through the incubating 6

sample to the surface marking the fact that gas has been produced. Visible interpretation of the test is therefore simple (D thimble down is negative, D thimble up and floating is positive). There is a relationship existing between population and the time lag to the vertical relocation of the D thimble. Here, the greater the size of the population then the sooner the D thimble will elevate. In the statistical evaluation of the COLI-BART test method, comparisons were made between the time lag to elevation (TL) and the population sizes for various pure (ATCC) cultured and natural populations of coliform bacteria. Two formats of the biodetectors have been developed. The first system detects the total (general) coliform population and is known as the T-COLI BART. The medium includes beef extract, peptone, lactose, tryptone/tryptose, dipotassium hydrogen phosphate, potassium dihydrogen phosphate, sodium chloride, sodium lauryl sulfate and bromocresol purple. The second system is more specifically designed to detect the fecal coliform bacteria uses brilliant bile green broth (F-COLI BART). This culture medium restricts the range of coliform bacteria able to ferment and produce gas to the target group comprised mainly of Escherichia coli. All of the tests were performed at temperatures of 37±1.0 o C using a baseline of 100 ml of double strength medium which was diluted to the correct final strength with a 100 ml water sample which had been appropriately prepared (e.g., as a natural water sample or a suspension of a known diluent of a 24 hr broth culture of a specific bacterial strain). The cultural conditions would be similar to the standard MPN techniques bearing similar potentials for false negatives (failure to detect presence) and false positives (detected activity not confirmed for the targeted organism). The provisional limits and conditions for testing using the COLI-BART system would be: Threshold detection limit (minimal): 1 cell per 100ml of sample Threshold detection limit (maximal): 10 billion cells per 100ml of sample Time length limitation: Total coliform bacteria, 20 hours Fecal coliform bacteria, 27 hours E. coli bacteria, 24 hours Water sample size: 100±2ml Water sample clarity: < 40 N.T.U. Water sample density: <1.02 @ 15 o C Water sample salinity: <3.7g/kg @ 15 o C Water sample storage: following the standard conditions laid down in the Standard Methods for the Examination of Water and Waste Water. Container: PET bottle with a capacity of 250ml Medium volume: 75±1ml ml of x2.34 strength sterile liquid selective medium Sample volume: 100±2ml Total Liquid volume: 175±3ml Height of detection light pathway to detect thimble elevation: 63mm to midpoint lateral 7

Height of gas detection thimble: Entrapment volume within the thimble: 3.6ml Volume of gas required to elevate the thimble in distilled water: 0.4±0.01ml Density of the gas detection thimble: 1.08 Ratio of slat area with closed area in lower part of the gas detection thimble: 1: 1.5 The patented components in the system are: COLI-BART system has a basic set-up that is similar to the BOD-BART system in that here are two components that form the system. These are described below as a tester (that performs an individual test) and the reader (that is able to monitor a multitude of testers concurrently and generate a predicted population for the samples under test). COLI-BART tester is a 250ml plastic sterile bottle that contains a dense plastic D thimble and 75ml of the appropriate selective culture medium that is recognized for the purposes of detecting either total coliform bacteria (T-COLI BART) or fecal coliform bacteria (F-COLI BART). The tester is sleeved in a clear plastic pouch to provide additional protection to the user, allow improved shipment of the final tester for disposal, and help to ensure that the culture within the positive finished tester could be securely taken to the laboratory for confirmatory testing and disposal. The technology for manufacturing the tester already exists. This would include aseptic procedures, clean air dispensing and adequate quality management. Outstanding issues relate to the form and functionality of the D thimble. The D thimble (version II) could also function in brackish (coastal) waters by the use of a density modifying attachment. COLI-BART reader primarily consists of the necessary hardware / software to be able to detect the upwards movement of the D thimble in the COLI-BART tester being incubated. This detection is achieved by the blocking off of an infra-red light pathway through the testers liquid medium by the D thimble once it has elevated by more than 4mm from its sunken position in the tester. Quantitative information is obtained by correlating the time lag to the thimble elevation given that the larger the population of coliform bacteria then the more rapid would be the generation of gases to cause the thimble to elevate. Product System Potential major users of coliform detection system is listed in Table Two. Table Two Potential User Groups for the Various BART Products Potable Water Waste Water Treatment Oil & Gas Industry Chemical Industry Environ. Industry Agriculture COLI-BART *** *** * ** *** *** Note: The potential for each industry listed in each column is shown by the number of asterisks defined below: 8

*** Could become an essential routine part of the testing / diagnostic procedure ** Under some circumstances could become a part of the routine testing but would be a valuable diagnostic tool under some conditions * May have considerable value under some limited conditions Results and Discussion At all stages, spreadplate analyses of the diluents and natural samples were used to obtain duplicate population assessments. Regression correlations (R 2 ) were calculated by linear regression analysis of the TL obtained to the populations of coliform bacteria determined by spreadplate analysis. In all cases the TL was determined as the time delay to the blocking of a near infra-red light pathway set 4 mm below the surface of the cultured medium. Monitoring for the TL event (elevation of the D thimble) was controlled by a microprocessor-driven software program. Three ATCC strains of Escherichia coli (ATCC 25922, 29839 and 11229) were subjected to triplicate trials using tenfold dilutions of the cell suspension down from 10 7 to 10-2 cells /100mL. Using the T-COLI BART protocol, triplicate analyses were performed. The correlations between the population density in the samples and the TL obtained for each individual trial were very good (R 2 values of 0.89, 0.96, 0.83, 0.81, 0.86, 0.96, 0.93, 0.83, 0.91). Pooling the data into a single scatter gram gave a reduced regression correlation with an R 2 value of 0.66 due to the variations inherent between the different strains of Escherichia coli used. The regression correlation equation for the pooled data was calculated to be: y = 0.5 x + 10.5 Where x is the time delay to D thimble elevation expressed in TL recorded as hours of delay. Y is the log population of coliform bacteria. The standard deviation around the y value was + 1.71. In general, higher levels of correlations were observed between the F-Coli BART tests using the same strains of Escherichia coli. Respective R 2 values were: 0.98, 0.96, 0.91, 1.00, 0.94, 0.60, 0.89, 0.88 and 0.83 to give a total scatter gram R 2 of 0.74 (compared to 0.66 for the T- protocol). The regression correlation equation for the pooled data was calculated to be: y = 0.47 x + 10.4 where x is the time delay to D thimble elevation expressed in TL recorded as hours of delay while y is the log population of coliform bacteria. The standard deviation around the y value was 1.91. While the incubation time of 24 hours appears to be more than adequate to recover low populations of coliform bacteria using the T- protocol, there has been incidents (one in every six occasions) where very low populations of coliform bacteria (<4 cells/100 ml) fail to cause D thimble in 24 hours in the F- protocol. It is recommended that for the F-COLI BART test procedure, the incubation period should routinely be extended to 30 hours in order to recover these very low coliform samples as positive. In practise, it has generally been noticed that all D thimble elevation in these very low population conditions occurs by 28 hours. The recommendation of 30 hours for incubation provides an additional two hour "safety" window. 9

Investigations in the summer of 2002 at a Saskatchewan POTW were focussed on the comparison of the TL for the T- and F- COLI-BART and the bacteriological data generated by the POTW using spread plate agar techniques. Typical variability for a single duplicate run is shown in Table Three and a summary of the variability for samples drawn through the third week of July (15 22, 2002) is given in Table Four. Table Three Single day variability in TL (seconds) and bacterial plate counts (Bact. colony forming units/100ml) for tertiary influent (TI), tertiary effluent (TE) and final effluent (FE) COLI-BART system Bact. Bact. Bact. TI TE FE TI TE FE TC 34648 39338 49135 460000 120000 200 33839 38868 47064 430000 90000 100 Average 34244 39103 48100 445000 105000 150 572 332 1464 21213 21213 71 1.7% 0.8% 3.0% 4.8% 20.2% 47.1% FC 37646 40417 58233 25000 6200 32 36933 43784 53328 26000 4500 24 Average 37290 42101 55781 25500 5350 28 Stand. Dev. 504 2381 3468 707 1202 6 % variability 1.4% 5.7% 6.2% 2.8% 22.5% 20.2% Table Four Summary of data comparisons for the T- and F- COLI BART tests against Bacteriological Counts (cfu/100ml) for POTW tertiary influent (TI), tertiary effluent (TE) and final effluent (FE) unit TI (T-Coli) TE (T-Coli) FE (T-Coli) TI (cfu/100ml) TE (cfu/100ml) FE (cfu/100ml) Average 35,802 43,254 50,989 330,000 54,400 625 TL Av. 994 3,154 4,416 21,213 17,678 106 Stand. Dev. % var. 2.8% 3.6% 8.5% 7.4% 27.3% 24.6% (F- COLI) (F- COLI) (F- COLI) Fecal bacteria Fecal bacteria Fecal bacteria Average 44,309 50,482 61,491 14,500 937 28 TL Av. 805 3,240 2,974 701 937 3 Stand. Dev. % var. 1.8% 11.6% 14.6% 2.8% 32.6% 50.7% 10

*Note: the percentage variability is an average of the variability determined from the means and standard deviations for all of the relevant trials. An inverse relationship between the TL (getting longer) as the bacterial population detected gets smaller through the treatment process (TI TE FE). Going downstream through the treatment variability increased from (T- : F- COLI BART TL) from 1.4 1.7% to 3.0 6.2% while, for the bacterial counts the variability was much more pronounced increasing from 4.8 47.1% and 2.8 20.2% for the total and fecal coliform bacteria. A similar trend was established for the seven day study period (given in Table Four) where the variability was (for the T- and F- COLI BART) 2.8 1.8% rising to 8.5 14.6% while the bacterial counts using the appropriate spreadplate techniques showed a variability rising from 3.6 11.6% to 24.6 and 50.7% from the TI to the FE. These differences are also shown in Figure One. Figure One Mean % variability for the four methods applied to the tertiary influent (1), Tertiary effluent (2) and final effluent (3). Variability of COLI-BART against spreadplate 60.0% 50.0% % varaition (sd/average) 40.0% 30.0% 20.0% T-Coli-BART Total Coli bacti F-Coli-BART Fecal Coli bacti 10.0% 0.0% 1 2 3 1 - TI, 2 - TE, 3 - FE The prime feature used in the MPN test is the determination of presence by the evolution of gases by fermentation during the test procedure. In the standard MPN using the small Durham s tube the total gas containment possible is 0.98 ml over a 50mm length meaning that each mm of gas entrapped within the tube would represent 0.02mL of gas collected. If the Durham s tube were to be used during an MPN procedure then the void volume (retained within the tube) ratio to the total volume (10mL) would be 1: 11

0.098 which means that much of the fermentation gases generated outside of the tube would not be captured leaving the efficiency of the system at collecting gas at less than 10%. In the COLI-BART system the D thimble has a constricted volume of 19.8 ml and a contained volume (total gas entrapment) of 5.29 ml. Elevation of the D thimble occurs when 0.4 ml of gas has been collected and the liquid medium has been displaced from the top 0.9 mm under the top plate of the D thimble. For the D thimble then the void volume (retained within the tube) ratio to the total volume (10mL) would be 1: 0.12which means that much of the fermentation gases generated outside of the tube would not be captured leaving the efficiency of the system at collecting gas at 12 % when submerged but this would increase as the D thimble elevates to the surface of the medium. This may be witnessed by the ongoing movement of the D thimble out of the liquid medium (Table Five) after a positive detection has been declared. Table Five Elevation of the D thimble out of the liquid medium after a COLI-BART Has been declared positive. Volume gas Entrapped (ml) T-COLI Volume gas Entrapped (ml) F-COLI Height (cm) T-COLI thimble Height (cm) F-COLI thimble TI 2.65 2.9 0.9 cm 1.0 cm 2.65 2.9 0.9 cm 1.0 cm TE 3.15 3.4 1.1 cm 1.2 cm 3.4 3.15 1.2 cm 1.1 cm FE 3.4 1.9 1.2 cm 0.6 cm 3.15 1.65 1.1 cm 0.5 cm *Note: the maximum contained volume in the D thimble would be 5.3 ml before gases would bleed out of the underside of the device. The data establishes the potential for an effective quantitative link between the traditional MPN test for the coliform bacteria that are present in waters and waste waters and the proposed possible population number (PPN) that would be based on the time lag preceding the evolution of sufficient gases to trigger the elevation of the D thimble. Such a COLI-BART system-based biotechnology would allow the use of a single vessel to determine quantitatively the population of the targeted coliform bacteria in the water or waste water sample. In this PPN system the major diversion from the standard protocols would be the replacement of the Durham s tube with the D thimble and then the quantification of the population using the time lag measured in seconds from the start of the test to the elevation of the D thimble to the medium/air interface. 12

Conclusions The patented COLI-BART system when applied to comparable studies against pure culture and natural samples showed a relationship existed between the time lag to the population size (from the pure culture studies) and also indicated a better precision in the quantitative determination of coliform bacteria populations in water and waste water samples. There are clear advantages in the proposed system in that it would be simpler to operate, have the potential to contain re-usable components (vessel, D thimble and cap) and would develop data that would have precision. Further testing is planned now to compare this system to the other commercially available methods for the determination of the presence of coliform bacteria. Acknowledgements The authors wish to acknowledge the very significant help that was received through the supply of samples, data and encouragement that was received from the many operators at the various sewage treatment plants who provided samples and data Also the financial assistance of the National Research Council of Canada through the Research Development Assistance program operated by the Industrial Research Assistance Program is acknowledged. References Alford, G. A., Rogers, W. and Cullimore, D.R., 1988. Well Cleaning Methods and Apparatus. U.S. Patent Number: 4765410. Buchanan, R.E., and N.E. Gibbons, 1974. Bergey s Manual of Determinative Bacteriology. R.E. Buchanan and N.E. Gibbons, Co-ed., The Williams & Wilkins Company, Baltimore, MD. Carter, Charles F., and Alice L. Smith, 1953. Microbiology and Pathology. Fifth Edition. The C.V. Mosby Company, St. Louis, MI. Cullimore, D.R., 1993. Practical Manual of Groundwater Microbiology. Lewis Publishers, Chelsea, MI. Cullimore, D.R., 1999. Microbiology of Well Biofouling. Sustainable Water Well Series, D.R. Cullimore, Series ed., Lewis Publishing/CRC Press, Boca Raton, Florida. Cullimore, D.R., 2000. Practical Atlas for Bacterial Identification. D.R. Cullimore, Lewis Publishing/CRC Press, Boca Raton, Florida. Cullimore, D.R. and Alford, G.A., 1990. Method and Apparatus Producing Analytic Culture. U.S. Patent Number: 4,906,566. 13

Cullimore, D.R. and Johnston, L., 2000. Microbial Issues in Water Using an Innovative Detection System. Water Conditioning & Purification Magazine. November Issue. Droycon Bioconcepts Incorporated, 2000. BART User Manual, 2000 Edition. Inhouse Publication. Hattori, T., 1988. The Viable Count. Science Technology Publishers, Madison Wisconsin. Park, William Hallock and Anna W. Williams, 1910. Pathogenic Micro-organisms including Bacteria and Protozoa. Fourth Edition. Lea & Febiger, New York, NY. Thompson, LaVerne Ruth, 1954. Introduction to Microorganisms. Third Edition. W.B. Saunders Company, Philadelphia, PA. Vinebrooke, R. D. and Cullimore, D.R., 1998. Natural Organic Matter and the Bound- Water Concept in Aquatic Ecosystems. Reviews of Environmental Contamination and Toxicology. Volume 155. Zapffe, Fred C. M.D., 1903. Bacteriology, A Manual for Students and Practitioners. Series ed., Bern B. Gallaudet, M.D. Lea Brother & Co, New York. 14