The aim of this study was to estimate the long-term (month-to-month) between-dog, within-dog and analytical components of variance for fasting plasma glucose and serum fructosamine in healthy dogs to assess the usefulness of a single measurement of these analytes in a single dog. Fasting plasma glucose and serum fructosamine were measured in blood samples collected every month for 9 months from 23 clinically healthy dogs, and the results were subjected to nested analysis of variance. The between-dog variation, the within-dog variation, and the analytical variation were 3.8%, 9.5% and 3.7%, respectively, for plasma glucose and 4.2%, 11.1% and 2.8%, respectively, for serum fructosamine. The maximum allowable analytical imprecision, analytical inaccuracy and difference between analytical methods were 4.8%, 2.6% and 3.2%, respectively, for plasma glucose and 5.6%, 3.0% and 3.7%, respectively, for serum fructosamine. The index of individuality, 2.7 for both analytes, indicated that the test results from single dogs can be compared usefully to the corresponding population-based reference intervals. The number of samples required to estimate the true individual mean value +/-5% for a single dog was 16 for fasting plasma glucose and 20 for serum fructosamine. The one- and two-sided critical differences expressing the difference needed for two serial results from the same dog to be significantly different at a 5% level was 24% and 28%, respectively, for plasma glucose and 27% and 32%, respectively, for serum fructosamine.