Skip to main content
. 2013 Jul 23;8(7):e70157. doi: 10.1371/journal.pone.0070157

Figure 2. Relationship between the prevalence of A. astaci (in %; estimated as the proportion of individuals testing positive) in analysed French signal crayfish populations and the average pathogen load (expressed as log-transformed PFU-values) detected in infected individuals (with agent level A2 or higher) from each population.

Figure 2

The equation characterising the model estimated by logistic regression (dashed line) is given in the upper right corner; removal of two outlier populations with highest average pathogen load (indicated by empty circles) did not change the model substantially (dotted line, equation not shown). Note that when a quasibinomial instead of binomial distribution of errors is used (due to overdispersion in data), the increasing trend is non-significant.