Vörös A. szerk.: Fragmenta Mineralogica Et Palaentologica 11. 1983. (Budapest, 1983)

Afn Mg A • +• X o f/rá/yréf fp/nk) aq/dázó K/rá/yréf Tödére f high Mn) Csehvár 70 Fe Fig. 7 Chemical variations of different garnet crystals in the triangle pyrope-almandine-spessartine. Arrows indicate chemical changes proceeding from the center to the margin Perhaps the most important difficulty in interpreting garnet megacrysts as cognate is the fact that high p, t experiments have failed to produce garnets of appropriate composition from dy-normative andesitic liquids. Though garnet was a liquidus phase in the experiments of STERN & WYLLIE (1978), GREEN (1972) and GREEN & RINGWOOD (1968, 1968 a) between 9 to 36 kb pressure from 1400 to 800 C and under conditions ranging from anhydrous to water-saturated they contained considerably higher amounts of pyrope and grossular molecules than natural garnets. BROUSSE & al. (1972) called the attention to the fact, however, that the composition of the natural garnet megacrysts especially the low Mn and the relatively high Ti content are also unusual in metamorphic garnets. Mineral inclusions in garnet (mainly apatite, zircon, ilmenite, magnetite) observed at many places are not at variance with the magmatic origin. The high content of plagioclase and amphibole inclusions in some of the Hungarian garnets seems to be an uncommon feature. Electron probe analyses (Table 2) have proved that the composition of these minerals closely resembles that of the same minerals in the host rock (cf. NOSKE-FAZEKAS 1980, PANTO 1970). As the inclusion-rich garnets are intensively ruptured as well, especially in their outer zones (Plate III, Fig. 1) it seems justified to think that liquid penetrated along cracks and inclusion minerals have crystallized pushing apart particles of garnet eventually replacing them. Zoning patterns of Hungarian garnets seem relatively consistent within individual host rocks but very variable between localities just like in Slovakia (BROUSSE & al. 1 972). The strongly reversed zoning pattern of the Csehvár garnets closely resembles that of the garnet crystal G-302 of the English Lake District (cf. FITTON 1972, fig. 2) that was interpreted to have started to crystallize in a dacite and later in an andésite. In contrast, the zoning pattern of the Bajdázó Hill garnets cannot be compared to any of the known cases. It was argued by GILL (1981) that zoning pattern is a worthless tool in