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

distinguishing between xenocrystal and cognate garnets because normal, reverse and oscillatory zoning occur in both cases. However, we do wish to emphasize that all kinds of Hungarian, Slovakian (BROUSSE & al. 1 972) English (FITTON 1972) and Australian (GREEN & RINGWOOD 1968) garnet zoning patterns differ basically from the so-called "bell-type" zoning (where Mg and Fe increase sympathetically from the core to the border which is counterbalanced by the reverse change in Mn) considered as the most characteristic zoning type present in metamorphic garnets (BLACK 1 973, DUDLEY 1969, HOLLISTER 1966, ATHERTON & EDMUNDS 1966). If so, it should also be present among the zoning patterns of the garnet megacrysts. Since it is not the case, we think that the evidence from the zoning patterns speaks likely for a deep-seated magmatic origin of the garnet megacrysts. Summing up, we can say that even if the dilemma of magmatic versus accidental origin could not be solved largely due to lack of exact knowledge of the physico-chemical conditions during the precipitation of garnets from natural andesitic liquids, considerably more observations and arguments can be advanced to support a cognate origin for the garnet megacrysts as opposed to the metamorphic provenance. According to the experimental evidence, the megacrysts must have precipitated either in the lower crust or rather in the upper mantle. The sporadic and "accidental" occurrence of the garnet may be explained by the fact that in most cases the crystallization took place in much shallower magma chambers where garnet was no more stable as was proposed by FITTON (1 972). Alternatively it can be also assumed that most magmas containing early formed garnet crystals stopped on their way to the surface at shallower situated secondary magma chambers where garnets had disappeared due to instability. Deepseated origin followed by a stay at a shallower crustal magma chamber have also been ascertained for some other minerals in the Börzsöny Mts, especially the Godóvár clinopyroxenes should be mentioned in this regard (EMBEY-ISZTIN & NOSKE-FAZEKAS 1981). Garnets of Tóbércbánya These garnets differ from the preceding ones in several respects: First, in the nature of the host rock and mode of occurrence. Second, in textural features such as shape and inclusion content of the crystals. Third, chemical composition and zoning. Fourth, by the fact that individual grains in the same host rock show important variations in chemical composition and zoning. The higher almandine + spessartine and lower grossular content of the Tóbércbánya garnets make them more similar to some garnets originating from granite or aplite (cf. MIYASHJ.RO 1955). Considering the problem of genesis, one point is very clear, namely that in no case can these garnets be xenocrystal in origin since they occur in a late magmatic differenciated rock of restricted dimensions surrounded by a normal gabbro and diabase containing no trace of any garnet. Direct precipitation from a melt is not likely either, because garnet is a liquidus phase only at pressures corresponding to depths prevailing in the lower crust and upper mantle (KEESMANN & al. 1971, GREEN 1 972, GREEN & RINGWOOD 1968, 1968 a). In contrast, the gabbro along with its fine-grained garnetiferous albite granite nests seem to represent a high level intrusion. The diabase body ( 10 km long and several km wide) was emplaced just below the surface or perhaps at some places true lavas may also be present (SZENTPETERY 1 953). The smaller gabbroic masses such as Tóbércbánya and Tardosbánya are closely connected with the diabase body (in the quarry of Tardosbánya the two rocks are grading in each other) consequently they should be regarded as very shallow intrusive rocks. Since the above two possibilities can be excluded, we should assume some late magmatic kind of origin for the garnets. GREEN (1915) proposed that hydrothermal processes may be responsible for garnet genesis. In our case, however, a metasomatic replacement origin seems to be more probable. First of all, in general the genesis of the irregular shaped small masses of aplitic granite itself may be explained in this way (GOODSPEED 1940, KING 1 948, PERRIN & ROUBAULT 1949, WELLS & BISHOP 1 954). In addition, the lack of euhedrality of the garnet crystals can be easily understood accepting this suggestion. Finally, the