Estimation of Staphylococcus aureus growth parameters from turbidity data: characterization of strain variation and comparison of methods

Turbidity methods offer possibilities for generating data required for addressing microorganism variability in risk modeling given that the results of these methods correspond to those of viable count methods. The objectives of this study were to identify the best approach for determining growth parameters based on turbidity data and use of a Bioscreen instrument and to characterize variability in growth parameters of 34 Staphylococcus aureus strains of different biotypes isolated from broiler carcasses. Growth parameters were estimated by fitting primary growth models to turbidity growth curves or to detection times of serially diluted cultures either directly or by using an analysis of variance (ANOVA) approach. The maximum specific growth rates in chicken broth at 17 degrees C estimated by time to detection methods were in good agreement with viable count estimates, whereas growth models (exponential and Richards) underestimated growth rates. Time to detection methods were selected for strain characterization. The variation of growth parameters among strains was best described by either the logistic or lognormal distribution, but definitive conclusions require a larger data set. The distribution of the physiological state parameter ranged from 0.01 to 0.92 and was not significantly different from a normal distribution. Strain variability was important, and the coefficient of variation of growth parameters was up to six times larger among strains than within strains. It is suggested to apply a time to detection (ANOVA) approach using turbidity measurements for convenient and accurate estimation of growth parameters. The results emphasize the need to consider implications of strain variability for predictive modeling and risk assessment.     

Estimation of Staphylococcus aureus Growth Parameters from Turbidity Data: Characterization of Strain Variation and Comparison of Methods

Turbidity methods offer possibilities for generating data required for addressing microorganism variability in risk modeling given that the results of these methods correspond to those of viable count methods. The objectives of this study were to identify the best approach for determining growth parameters based on turbidity data and use of a Bioscreen instrument and to characterize variability in growth parameters of 34 Staphylococcus aureus strains of different biotypes isolated from broiler carcasses. Growth parameters were estimated by fitting primary growth models to turbidity growth curves or to detection times of serially diluted cultures either directly or by using an analysis of variance (ANOVA) approach. The maximum specific growth rates in chicken broth at 17°C estimated by time to detection methods were in good agreement with viable count estimates, whereas growth models (exponential and Richards) underestimated growth rates. Time to detection methods were selected for strain characterization. The variation of growth parameters among strains was best described by either the logistic or lognormal distribution, but definitive conclusions require a larger data set. The distribution of the physiological state parameter ranged from 0.01 to 0.92 and was not significantly different from a normal distribution. Strain variability was important, and the coefficient of variation of growth parameters was up to six times larger among strains than within strains. It is suggested to apply a time to detection (ANOVA) approach using turbidity measurements for convenient and accurate estimation of growth parameters. The results emphasize the need to consider implications of strain variability for predictive modeling and risk assessment.     

Development of a microbial model for the combined effect of temperature and pH on spoilage of ground meat, and validation of the model under dynamic temperature conditions

The changes in microbial flora and sensory characteristics of fresh ground meat (beef and pork) with pH values ranging from 5.34 to 6.13 were monitored at different isothermal storage temperatures (0 to 20 degrees C) under aerobic conditions. At all conditions tested, pseudomonads were the predominant bacteria, followed by Brochothrix thermosphacta, while the other members of the microbial association (e.g., lactic acid bacteria and Enterobacteriaceae) remained at lower levels. The results from microbiological and sensory analysis showed that changes in pseudomonad populations followed closely sensory changes during storage and could be used as a good index for spoilage of aerobically stored ground meat. The kinetic parameters (maximum specific growth rate [mu(max)] and the duration of lag phase [lambda]) of the spoilage bacteria were modeled by using a modified Arrhenius equation for the combined effect of temperature and pH. Meat pH affected growth of all spoilage bacteria except that of lactic acid bacteria. The "adaptation work," characterized by the product of mu(max) and lambda(mu(max) x lambda) was found to be unaffected by temperature for all tested bacteria but was affected by pH for pseudomonads and B. thermosphacta. For the latter bacteria, a negative linear correlation between ln(mu(max) x lambda) and meat pH was observed. The developed models were further validated under dynamic temperature conditions using different fluctuating temperatures. Graphical comparison between predicted and observed growth and the examination of the relative errors of predictions showed that the model predicted satisfactorily growth under dynamic conditions. Predicted shelf life based on pseudomonads growth was slightly shorter than shelf life observed by sensory analysis with a mean difference of 13.1%. The present study provides a "ready-to-use," well-validated model for predicting spoilage of aerobically stored ground meat. The use of the model by the meat industry can lead to effective management systems for the optimization of meat quality.     

Sand administration as an instrument for biofilm control of Pseudomonas putida ATCC 11172 in chemostat cultures

Pseudomonas putida ATCC 11172 was grown in chemostat on L-asparagine or phenol as the sole, limiting carbon and energy source. The growth characteristics of a culture where a biofilm was present, were compared with one where the biofilm was strongly reduced by the grinding and shearing effect of sand suspended in the culture. In the presence of the intact biofilm, the curve of steady-state biomass versus dilution rate diverged greatly from the theoretical pattern predicted by conventional chemostat models. The sand strongly retarded the biofilm formation and to a high degree restored the shape of the biomass versus dilution rate curve to a more conventional pattern. The maximum specific growth rate (mu(max)) could not be calculated from the biofilm cultures. However using the sand cultures, mu(max) was determined to 0.64 h(-1) with L-asparagine as the carbon source and 0.49 h(-1) with phenol which compare favorably with the respective mu(max) values calculated from batch cultures.Incorporation of sand into strongly agitated cultures is recommended as an efficient and simple means of controlling biofilm formation in continuous cultures. The method may enable the gathering of basic kinetic data difficult to obtain in the presence of biofilm.     

Growth kinetics of Escherichia coli with galactose and several other sugars in carbon-limited chemostat culture

Kinetic models for microbial growth describe the specific growth rate (mu) as a function of the concentration of the growth-limiting nutrient (s) and a set of parameters. A typical example is the model proposed by Monod, where mu is related to s using substrate affinity (Ks) and the maximum specific growth rate (mu max). The preferred method to determine such parameters is to grow microorganisms in continuous culture and to measure the concentration of the growth-limiting substrate as a function of the dilution rate. However, owing to the lack of analytical methods to quantify sugars in the microgram per litre range, it has not been possible to investigate the growth kinetics of Escherichia coli in chemostat culture. Using an HPLC method able to determine steady-state concentrations of reducing sugars, we previously have shown that the Monod model adequately describes glucose-limited growth of E. coli ML30. This has not been confirmed for any other sugar. Therefore, we carried out a similar study with galactose and found steady-state concentrations between 18 and 840 micrograms.L-1 for dilution rates between 0.2 and 0.8.h-1, respectively. With these data the parameters of several models giving the specific growth rate as a function of the substrate concentration were estimated by nonlinear parameter estimation, and subsequently, the models were evaluated statistically. From all equations tested, the Monod model described the data best. The parameters for galactose utilisation were mu max = 0.75.h-1 and Ks = 67 micrograms.L-1. The results indicated that accurate Ks values can be estimated from a limited set of steady-state data when employing mu max measured during balanced growth in batch culture. This simplified procedure was applied for maltose, ribose, and fructose. For growth of E. coli with these sugars, mu max and Ks were for maltose 0.87.h-1, 100 micrograms.L-1; for ribose 0.57.h-1, 132 micrograms.L-1, and for fructose 0.70.h-1, 125 micrograms.L-1.     

Growth physiology and competitive interaction of obligately chemolithoautotrophic,haloalkaliphilic, sulfur-oxidizing bacteria from soda lakes

Two different groups of haloalkaliphilic, obligately autotrophic, sulfur-oxidizing bacteria belonging to the genera Thioalkalimicrobium and Thioalkalivibrio have recently been discovered in highly alkaline and saline soda lakes. To understand response to their extreme environment and different occurrence in soda lakes, the growth kinetics and competitive behavior of several representatives have been characterized in detail using batch and pH-controlled continuous cultivation. The bacteria belong to the true alkaliphiles, growing within the pH range 7.5-10.6 with maximum growth rate and maximum growth yield at pH 9.5-10. On the basis of their response to salt content, three groups can be identified. All the Thioalkalimicrobium strains and some of the Thioalkalivibrio strains belonged to the moderate halophiles. Some of the Thioalkalivibrio strains from hypersaline soda lakes were extremely salt-tolerant and capable of growth in saturated soda brines. The Thioalkalimicrobium strains demonstrated relatively high specific growth rates, low growth yield, high maintenance, and extremely high rates of thiosulfate and sulfide oxidation. In contrast, the Thioalkalivibrio strains, in general, were slow-growing, high-yield organisms with lower maintenance and much lower rates of oxidation of sulfide and thiosulfate. Moreover, the latter survived starvation much better than Thioalkalimicrobium. Different growth characteristics and salt resistance appear to determine the outcome of the enrichment cultures from different soda lakes: Thioalkalimicrobium dominated in the enrichments with freshly obtained samples from diluted soda lakes at low-medium salinity, while Thioalkalivibrio was the predominant organism in enrichments from aged samples and at hypersaline conditions. In mixed thiosulfate-limited chemostat cultures at low salinity, Thioalkalimicrobium strains (mu(max)=0.33 h(-1)) out-competed Thioalkalivibrio strains (mu(max)=0.15 h(-1)) at D>0.02 h(-1). The overall results suggest that Thioalkalimicrobium and Thioalkalivibrio represent two different ecological strategies.     

Estimation of temperature dependent growth rate and lag time of Listeria monocytogenes by optical density measurements.

An automated turbidimetric system, Bioscreen C, was used to monitor growth of ten strains of Listeria monocytogenes at different temperatures. Several methods for estimation of maximum specific growth rate (mu(max)) and lag time (lag) from turbidimetric data were compared to values estimated from viable count data. By using a calibration factor, reliable estimations of mu(max) could be obtained from turbidimetric measurements. On the other hand, accurate estimations of lag required some viable count data.     

Specific growth rate determines the sensitivity of Escherichia coli to thermal, UVA, and solar disinfection

Knowledge about the sensitivity of the test organism is essential for the evaluation of any disinfection method. In this work we show that sensitivity of Escherichia coli MG1655 to three physical stresses (mild heat, UVA light, and sunlight) that are relevant in the disinfection of drinking water with solar radiation is determined by the specific growth rate of the culture. Batch- and chemostat-cultivated cells from cultures with similar specific growth rates showed similar stress sensitivities. Generally, fast-growing cells were more sensitive to the stresses than slow-growing cells. For example, slow-growing chemostat-cultivated cells (D = 0.08 h(-1)) and stationary-phase bacteria from batch culture that were exposed to mild heat had very similar T(90) (time until 90% of the population is inactivated) values (T(90, chemostat) = 2.66 h; T(90, batch) = 2.62 h), whereas T(90) for cells growing at a mu of 0.9 h(-1) was 0.2 h. We present evidence that the stress sensitivity of E. coli is correlated with the intracellular level of the alternative sigma factor RpoS. This is also supported by the fact that E. coli rpoS mutant cells were more stress sensitive than the parent strain by factors of 4.9 (mild heat), 5.3 (UVA light), and 4.1 (sunlight). Furthermore, modeling of inactivation curves with GInaFiT revealed that the shape of inactivation curves changed depending on the specific growth rate. Inactivation curves of cells from fast-growing cultures (mu = 1.0 h(-1)) that were irradiated with UVA light showed a tailing effect, while for slow-growing cultures (mu = 0.3 h(-1)), inactivation curves with shoulders were obtained. Our findings emphasize the need for accurate reporting of specific growth rates and detailed culture conditions in disinfection studies to allow comparison of data from different studies and laboratories and sound interpretation of the data obtained.     

Visual pigments and optical habitats of surfperch (Embiotocidae) in the California kelp forest

We studied the optical microhabitat use and visual pigment variation among a group of closely related teleosts (surfperch: Embiotocidae) living along the nearshore central California coast. We employed a diver-operated spectroradiometer to record the optical microhabitat use of eight surfperch species in Monterey Bay. and microspectrophotometry to measure visual pigment absorbance for nine surfperch species. Species were dichromatic with mixtures of A1- and A2-based visual pigments exhibiting extensive maximum absorbance (lambda(max)) variation across species: 455-482 nm for SWS cones and 527-546 nm for LWS cones. Interspecific variation in sidewelling irradiance measurements (mean lambdaFmaxs) significantly accounted for 63% of the variation in surfperch LWS visual pigments and 83% of the interspecific variation in SWS visual pigments using a phylogenetically-corrected regression technique. Optimality models for maximizing relative photon capture of background radiance demonstrate that the LWS cone lambda(max) values are tuned for maximizing photon capture of the species-specific horizontal visual field, while the SWS cone lambda(max), are well offset from the dominant background radiance. This study is one of the first to demonstrate species-specific differences in habitat usage at microhabitat scales accounting for differences in photoreceptor peak absorbance among closely related, sympatric species.     

Modeling the lag time of Listeria monocytogenes from viable count enumeration and optical density data

The following two factors significantly influence estimates of the maximum specific growth rate ( micro (max)) and the lag-phase duration (lambda): (i) the technique used to monitor bacterial growth and (ii) the model fitted to estimate parameters. In this study, nine strains of Listeria monocytogenes were monitored simultaneously by optical density (OD) analysis and by viable count enumeration (VCE) analysis. Four usual growth models were fitted to our data, and estimates of growth parameters were compared from one model to another and from one monitoring technique to another. Our results show that growth parameter estimates depended on the model used to fit data, whereas there were no systematic variations in the estimates of micro (max) and lambda when the estimates were based on OD data instead of VCE data. By studying the evolution of OD and VCE simultaneously, we found that while log OD/VCE remained constant for some of our experiments, a visible linear increase occurred during the lag phase for other experiments. We developed a global model that fits both OD and VCE data. This model enabled us to detect for some of our strains an increase in OD during the lag phase. If not taken into account, this phenomenon may lead to an underestimate of lambda.     

Kinetic comparison of seven strains of 2,4-dichlorophenoxyacetic acid-degrading bacteria.

Seven strains of 2,4-dichlorophenoxyacetic acid-degrading bacteria, including Pseudomonas, Alcaligenes, and Bordetella spp., were compared on the basis of growth kinetics. Estimates of maximum growth rate (mu max, k1) and half-saturation growth constant (Ks, k3) were obtained by fitting substrate depletion curves to a four-parameter version of the integrated Monod equation. Estimates of Ks ranged from 2.2 micrograms/ml (10 microM) to 33.8 micrograms/ml (154 microM), and estimates of mu max ranged from 0.20 h-1 (Td = 3.5 h) to 0.32 h-1 (Td = 2.2 h). Estimates of mu max, but not Ks, were affected by changes in initial inoculum density. Maximum growth rates (mu max) were also estimated from turbidity measurements. They ranged from 0.10 h-1 (Td = 6.9 h) to 1.0 h-1 (Td = 0.7 h). There was no correlation between estimates of mu max derived from substrate depletion curves and those derived from turbidity measurements (P = 0.20).     

Kinetic comparison of seven strains of 2,4-dichlorophenoxyacetic acid-degrading bacteria.

Seven strains of 2,4-dichlorophenoxyacetic acid-degrading bacteria, including Pseudomonas, Alcaligenes, and Bordetella spp., were compared on the basis of growth kinetics. Estimates of maximum growth rate (mu max, k1) and half-saturation growth constant (Ks, k3) were obtained by fitting substrate depletion curves to a four-parameter version of the integrated Monod equation. Estimates of Ks ranged from 2.2 micrograms/ml (10 microM) to 33.8 micrograms/ml (154 microM), and estimates of mu max ranged from 0.20 h-1 (Td = 3.5 h) to 0.32 h-1 (Td = 2.2 h). Estimates of mu max, but not Ks, were affected by changes in initial inoculum density. Maximum growth rates (mu max) were also estimated from turbidity measurements. They ranged from 0.10 h-1 (Td = 6.9 h) to 1.0 h-1 (Td = 0.7 h). There was no correlation between estimates of mu max derived from substrate depletion curves and those derived from turbidity measurements (P = 0.20).     

Gene duplication is an evolutionary mechanism for expanding spectral diversity in the long-wavelength photopigments of butterflies

Butterfly long-wavelength (L) photopigments are interesting for comparative studies of adaptive evolution because of the tremendous phenotypic variation that exists in their wavelength of peak absorbance (lambda(max) value). Here we present a comprehensive survey of L photopigment variation by measuring lambda(max) in 12 nymphalid and 1 riodinid species using epi-microspectrophotometry. Together with previous data, we find that L photopigment lambda(max) varies from 510-565 nm in 22 nymphalids, with an even broader 505- to 600-nm range in riodinids. We then surveyed the L opsin genes for which lambda(max) values are available as well as from related taxa and found 2 instances of L opsin gene duplication within nymphalids, in Hermeuptychia hermes and Amathusia phidippus, and 1 instance within riodinids, in the metalmark butterfly Apodemia mormo. Using maximum parsimony and maximum likelihood ancestral state reconstructions to map the evolution of spectral shifts within the L photopigments of nymphalids, we estimate the ancestral pigment had a lambda(max) = 540 nm +/- 10 nm standard error and that blueshifts in wavelength have occurred at least 4 times within the family. We used ancestral state reconstructions to investigate the importance of several amino acid substitutions (Ile17Met, Ala64Ser, Asn70Ser, and Ser137Ala) previously shown to have evolved under positive selection that are correlated with blue spectral shifts. These reconstructions suggest that the Ala64Ser substitution has indeed occurred along the newly identified blueshifted L photopigment lineages. Substitutions at the other 3 sites may also be involved in the functional diversification of L photopigments. Our data strongly suggest that there are limits to the evolution of L photopigment spectral shifts among species with only one L opsin gene and that opsin gene duplication broadens the potential range of lambda(max) values.     

Modeling the growth of Enterococcus faecium in bologna sausage.

A study to set up mathematical models which allow the prediction of Enterococcus faecium growth in bologna sausage (mortadella) was carried out. Growth curves were obtained at different temperatures (5, 6, 12, 15, 25, 32, 35, 37, 42, 46, 50, 52, and 55 degrees C). The Gompertz and logistic models, modified by Zwietering, were found to fit with the representation of experimental curves. The variations of the parameters A (i.e., the asymptotic value reached by the relative population during the stationary growth phase), mu m (i.e., the maximum specific growth rate during the exponential growth phase), and lambda (i.e., the lag time) with temperature were then modeled. The variation of A with temperature can be described by an empirical polynomial model, whereas the variation of mu m and lambda can be described by the Ratkowsky model modified by Zwietering and the Adair model, respectively. Data processing of these models has shown that the minimum growth temperature for E. faecium is 0.1 degrees C, the maximum growth temperature is 53.4 degrees C, and the optimal growth temperature is 42 to 45 degrees C.     

Production of the fimbrial adhesin 987P by enterotoxigenic Escherichia coli during growth under controlled conditions in a chemostat

The effects of growth conditions on the production of 987P fimbriae by the enterotoxigenic Escherichia coli strain 1592 were examined in steady state chemostat experiments at different specific growth rates. The amount of fimbriae produced by fimbriate cells (P+) was dependent on the specific growth rate (mu). Under aerobic growth conditions fimbriae production increased with higher mu values till mu = 0.40 h-1 and decreased again at mu values close to mu max (0.48 h-1). Under anaerobic growth conditions the maximal production was comparable to that under aerobic growth conditions, and was also maximal close to mu max (0.16 h-1). Phase variation, measured as the percentage of fimbriate cells in a particular population, was independent of mu. The composition of the growth medium influenced both phase variation and overall production of fimbriae. A shift from minimal to a complex medium induced a rapid reduction in the amount of fimbriae per P+ cell and a slower reduction in the percentage of P+ cells. A shift from complex to minimal medium resulted in an increase in the percentage of P+ cells and a constant amount of fimbriae per P+ cell. The frequency of the phase switch was calculated for different growth conditions. The frequency of the P+----P- switch between two steady states was 2.7 x 10(-2). In batch culture the frequency of the P(-)----P+ switch was minimally 2.9 x 10(-2). The results indicate that phase variation and the production of 987P fimbriae by fimbriate cells are under independent physiological control.     

Evaluation of different phenol hydroxylase-possessing phenol-degrading pseudomonads by kinetic parameters

Phenol-degrading pseudomonads possessing different phenol hydroxylases (PH) were evaluated by the values of apparent half-saturation constant for phenol-oxygenating activity (K ( S )), maximum specific growth rate (mu (max)), lag-time length (lambda), inhibition constant (K ( I )) and growth yield factor (Y ( X/S )). Strains of the same PH type showed similar kinetic parameters: single-component PH (sPH) harbouring strains had higher values of K ( S ) and lower values of mu (max) than the strains having multicomponent PH (mPH). However, the values of K ( I ) and the dependencies of the lag-time length on initial phenol concentration were strain-specific. The elevated ratio between specific activities of catechol 1,2-dioxygenase (C12O) and muconate cycloisomerase in sPH-strains caused irreversible accumulation of a high amount of exogenous cis,cis-muconate (CCM) which resulted in decreased Y ( X/S ) values. Co-presence of sPH and mPH genes did not give the strains PC16 and P69 any extra advantage and according to determined kinetic parameters only one PH was active during phenol degradation. At the same time simultaneous functioning of catechol ortho and meta cleavage pathways (strain PC20) resulted in higher mu (max) and Y ( X/S ) values. Evaluation of strains showed that the type of PH determined the efficiency of phenol degradation, whereas the tolerance to elevated phenol concentrations was strain-specific.     

Variation and modeling of the probability of plasmid loss as a function of growth rate of plasmid-bearing cells of Escherichia coli during continuous cultures

A large number of models concerning cultures of genetically engineered bacteria have been described. Among them, some are specifically adapted to continuous cultures and lead to the determination of two variables: (i) the difference in the specific growth rates between plasmid-carrying cell and plasmid-free cells (deltamu) and (ii) the frequency of plasmid loss by plasmid-containing cells (p(r)mu(+)). Until now, studies have been performed on the global expression p(r)mu(+) and deltamu, whose value during continuous assays have been supposed approximately constant (mean value) and not on separate values of both terms p(r) and mu(+), respectively, probability of plasmid loss and specific growth rate of the plasmid-carrying cells. So far these studies do not allow examination of the relationship between these two last parameters. Experimental results were obtained with Escherichia coli C600 galk (GAPDH), a genetically engineered strain that synthetizes an elevated quantity of glyceraldehyde-3-phosphate dehydrogenase (GAPDH). From data obtained during continuous cultures, it is shown that during an assay, deltamu, and p(r)mu(+) do not remain constant. An appropriate mathematical analysis of the expression of mu(-) (specific growth rate of the plasmid-free cells) and mu(+) has been built up. This allows the evaluation of the values of mu(+) and mu(-) during the continuous cultures carried out at different dilution rates. Values of p(r) have been calculated from these data. Indeed our results show that p(r) increases with mu(+). A modeling approach which allows correct simulation of this variation is also proposed. This model is derived from the Hill equation regarding cooperative binding of enzymic type reaction. (c) 1993 John Wiley & Sons, Inc.     

The evaluation of mixtures of yeast and potato extracts in growth media for biomass production of lactic cultures

The effectiveness of yeast extracts (YE) and potato extracts (PE) to promote growth of seven lactic cultures was evaluated by automated spectrophotometry (AS). Two aspects of the growth curve were analysed: (1) maximum biomass obtained (using ODmax) and (2) highest specific growth rate mu(max)) Eleven lots from the same PE-manufacturing process were examined for lot-to-lot variability. The ODmax values of three of the seven strains were significantly affected by lot source, but mu(max) was not significantly affected. The growth of bacteria was systematically lower in base medium containing 100% PE than in base medium containing 100% YE for both ODmax or mu(max) data, which could be related to the lower content in nitrogen-based compounds in PE. In AS assays, highest OD values for Lactobacillus casei EQ28, Lactobacillus rhamnosus R-011, Lactobacillus plantarum EQ12, and Streptococcus thermophilus R-083 were obtained with a mixture of PE and YE. Fermentations (2 L) were also carried out to determine the accuracy of AS to predict biomass levels obtained under fermentation trials. In these fermentations, replacement of 50% YE with PE was shown to enable good growth of S. thermophilus. With L. rhamnosus R-011, a high correlation (R2 = 0.95) was found between ODmax data obtained in the AS assays and that of the 2-L bioreactor when the same growth medium was used for both series of fermentations. However, AS was not as efficient when industrial media were used for the bioreactor assays. The relationship was still good for ODmax between AS data and that of the bioreactor data with L. rhamnosus R-011 in industrial LBS medium (R2 = 0.87), but was very poor with the S. thermophilus R-083 on Rosell #43 industrial medium (R2 = 0.33). Since PE cost 40% less than YE, there are strong economic advantages in considering such a partial replacement of YE by PE.     

Analytical procedure for use of conductance measurement to estimate Escherichia coli in shellfish

Assays were performed with a Malthus AT Microbiological Analyzer to define an analytical procedure to estimate Escherichia coli counts in live bivalve shellfish by conductance measurement. The growth conditions used (Malthus Coliform Broth at 44°C) were selective for E. coli, and interference was noted only when Klebsiella pneumoniae were at least 100 times as numerous as E. coli. Different sample preparation procedures and seeding conditions were tested to obtain good quality conductance curves. The best results were observed when: (a) meat and shell liquor were diluted 1 : 3 with tryptone salt water and homogenized in a Waring blender for 1 min at 15 000 rev min^-1; and (b) the inoculum was taken from the liquid phase of the homogenate 20 min after blending and mixed immediately with the culture medium. Detection parameter threshold values were adjusted (first difference 1.5 μS for the baseline and 3.5 μS for detection, second difference 0.2 μS to improve detection time reliability. The repeatability of conductance measurements was very good (S.D. as % response mean ranged from 1.9 to 3.3) with the protocol used.