In vivo bioluminescence imaging (BLI) has the advantages of high sensitivity and low background. By counting the number of photons emitted from a specimen, BLI can quantify biological events such as tumour growth, gene expression and drug response. The intensities and kinetics of the BL signal are affected by many factors and may confound the quantitative results acquired from consecutive imaging sessions or different specimens. We used three different mouse models of tumours to examine whether anaesthetics, positioning and tumour growth may affect the consistency of the BL signal. The results showed that BLI signal could be affected by different anaesthetics and repetitive positioning. Using the same anaesthetics produced consistent peak times, while other factors were held constant. However, as the tumours grew the peak times shifted and the time course of BL signals had different shapes, depending on the positioning of the mice. The data indicate that a carefully designed BLI experiment is required to generate optimal and consistent results.