The Miocene Nyalam leucogranites from the South Tibetan Himalaya consist of two-mica leucogranite and tourmaline leucogranite, in which tourmaline is an important constituent in nodular tourmaline-quartz segregations. Within the two-mica leucogranite these nodules display two basic morphologies: rounded (2mg-ro type) and radial or dendritic (2mg-ra type). The nodules in the tourmaline leucogranite share similar textural features with the 2mg-ro type tourmaline nodules, but without the leucocratic halo. Tourmaline crystals in the tourmaline leucogranite display overgrowth textures — most commonly in the form of homogeneous cores (Tg-I type) surrounded by oscillatory-zoned rims (Tg-II type). Based on the petrography, chemical and boron isotopic compositions of the various types of tourmaline, it is hypothesized that the 2mg-ro type tourmaline nodules crystallized during the transition from magmatic to hydrothermal evolution of the granite, while the 2mg-ra type tourmaline nodules likely formed during syn-magmatic crystallization from boron-rich melts. In this hypothesis, the Tg-I type tourmalines also formed during the transitional stage from late magmatic to early hydrothermal process. Tg-II type tourmalines were obviously formed after the Tg-I type tourmaline, and can be attributed to the mixing of juvenile fluids and volatiles coming from the wall rocks to the nodule in a post-magmatic hydrothermal environment where Tg-I type tourmaline has already crystallized. Thus, the origin of the tourmaline nodules is related to the different stages of syn-magmatic crystallization from a B-rich melt, or the subsequent transition from magmatic to hydrothermal environment, or crystallized in the post-magmatic hydrothermal fluids.