Recently we reported on the detailed localization of melatonin (and its receptors) in human salivary glands, revealing that serous cells are able to store and secrete melatonin into saliva. Since we found that type 2 diabetic patients display reduced melatonin content in saliva, our next step was to examine the presence of melatonin in salivary glands removed from type 2 diabetic subjects. The resulting data were compared with those previously obtained by identical procedures in non-diabetics, to establish if the diabetic status may affect melatonin distribution. Bioptic samples of diabetic parotid and submandibular glands were fixed, dehydrated, embedded in Epon Resin and processed to demonstrate melatonin reactivity by the immunogold staining method. The labeling density (expressed as the number of gold particles per μm /granule) and the percentage of melatonin-positive granules were assessed in diabetic samples. These values were compared with those in non-diabetic samples and differences were evaluated. In parotid and submandibular diabetic glands the reactivity for melatonin was specifically associated with secretory granules and small vesicles in serous cells. Melatonin reactivity was higher in parotid than in submandibular glands. Our data were in line with those obtained in our previous study on non-diabetic glands. Diabetic salivary glands showed a higher labeling density and a lower number of melatonin-positive granules compared to non-diabetic glands. Taken together, these data might explain the decreased salivary melatonin content and the associated oral problems observed in diabetics. Anat Rec, 301:711-716, 2018. © 2017 Wiley Periodicals, Inc.