When forensic pathologists and anthropologists have to deal with the evaluation of the post-mortem interval (PMI) in skeletal remains, luminol testing is frequently performed as a preliminary screening method. However, the repeatability of this test on the same bone, as well as comparative studies on different bones of the same individual, has never been performed. Therefore, with the aim of investigating the influence that different types of bones may exert on the response to the luminol test, the present study analysed three different skeletal elements (femoral diaphysis, vertebra and cranial vault), gathered from ten recent exhumed skeletons (all with a 20-year PMI). The analysis was performed twice on the same bone after 2months: the analysis at time 0 concerned the whole bone, whereas the second concerned only a part of the same bone taken during the first test (which already had been broken). The overall results showed different responses, depending on the type of bone and on the integrity of the samples. Negative results at the first analysis (6.6% out of the total of samples) are consistent with what is reported in the literature, whilst at the second analysis, the increase of about 20% of false-negative results highlights that the luminol test ought to be performed with caution in case of broken bones or elements which are taphonomically altered. Results have thus proven that the exposition to environmental agents might result in haemoglobin (Hb) loss, as detected even after only 2months. The study also focused on the crucial issue of the type of bone subjected to testing, remarking the suitability of the femoral diaphysis (100% of positive responses at the first analysis vs only 18% of false-negative results at the second test, corresponding to 5% of total false-negative results) as opposed to other bone elements that showed a low yield. In particular, the cranial vault gave poor results, with 40% of discrepancy between results from the two analyses, which suggests caution in choosing the type of bone sample to test. In conclusion, luminol testing should be used with caution on bones different from long bones or on non-intact bones.