Variability of Vairimorpha (=Nosema) ceranae infection level in individual honey bees and its implications on the pooled sample size

Abstract

Nosemosis is caused by two microsporidian species of the genus recently renamed Vairimorpha (=Nosema) that parasitize the midgut of Apis mellifera. Usually, pooled samples of 60 bees are used for diagnosis. Reported differences in the efficiency of treatments against Vairimorpha sp. led us to wonder if the sample size may be one of the causes of these unpredictable results. This study focuses on the distribution of the spore load of V. ceranae among forager bees within a colony using statistical approaches in relation to sample size. Spore counts per bee were used to describe the spore distribution and to test whether they fit different statistical distributions. Standard errors were calculated and two-sided 95% confidence intervals were constructed for the spore count means, using standard (Student’s t) and non-parametric bootstrap methods (Percentil and bias-corrected and accelerated-BCa). Bees with a high degree of infection appear in low percentage but are very influential on the mean value. Data showed a poor fit to all of the distributions evaluated. To obtain a relative standard error below 10% of the mean, pooled samples should have more than 1,500 bees. The BCa interval showed better coverage than others being acceptable starting at 100 data of individual foragers. Results show that, when using the traditional sample size (pooled samples n = 60), the accuracy for estimating the mean spore load is low in V. ceranae infected hives, regardless of the statistical method applied. Thus, this work could justify some erratic results reported for treatments with commercial products.