Monazite is a valuable geochronometer but is highly prone to react after its formation. Its alteration reactions must therefore be understood to assess its geochronological potential as well as the mobility of rare earth elements and actinides. We study contrasting monazite alteration textures in two different lithologies from Mt. Papuk (Slavonian Mts., Croatia) in order to constrain the timing of thermal events in this eastern exposure of the European Variscan belt, and explore the role of fluids and host-rock composition during monazite alteration. For a migmatite and peraluminous granite sample, we document the texture and composition of monazite-(Ce) and associated phases (xenotime-(Y), actinide-rich phases) and date them by EPMA and/or LA-ICP-MS. The texture observed in the quartz – feldspar ± biotite migmatite (Trešnjevica) shows a replacement of primary magmatic monazite by secondary (Th-poorer and LREE-richer) monazite, xenotime and Th–Si-rich phases, and is locally rimmed by apatite. The re-integrated composition of this texture is compatible with that of (Th- and Y-richer) primary monazite. Primary monazite domains yield a magmatic crystallization age of 351± 9 Ma while secondary domains are contaminated by initial Pb or give slightly younger Th–U–total Pb and ²⁰⁸Pb/²³²Th dates (~340 Ma). The texture found in the two-mica granite (Zvečevo) involves replacement by apatite and allanite, with some secondary (Th-poorer) monazite and minute uraninite or Th–Si-rich phases. Relicts of primary monazite grains record magmatic crystallization at 352±9 Ma, like primary xenotime domains. Secondary monazite contains initial Pb, secondary xenotime domains yield ²⁰⁶Pb/²³⁸U dates younger than 338 Ma and uraninite preserves Th–U–total Pb dates averaging at 332 Ma. The first texture indicates a “dry” decomposition that lacks evidence for the mediation of a fluid phase and efficiently trapped all elements (except Y) at the location of the monazite precursor. It is attributed to monazite re-equilibration (Th and Y+HREE loss) at subsolidus conditions. The second texture points to a “wet” replacement by hydrous minerals that needs the presence of fluid and led to REE mobility at least at the milli­metre scale. The contrast between both reactions is ascribed to sample composition, one being an anhydrous migmatite hardly affected by fluid infiltration and the other offering all necessary components for Ca and H₂O mobility. Both alteration reactions are correlated with a late Variscan thermal event at ca. 335 Ma. The latter could also be dated using actinide-rich phases, in which case tiny U-oxides appear more reliable than Th-silicates.