Comments on Some Misconceptions in Igneous and Experimental Petrology and Methodology: a Reply

rate’ of cpx and opx. Isobaric melting experiments show INTRODUCTION a > b, but the abyssal peridotite data suggest b > a, i.e. I recently showed (Niu, 1997) that proportions of residual opx would contribute more than cpx to the melt during minerals in abyssal peridotites (e.g. Dick, 1989; Johnson melting beneath ocean ridges. As abyssal peridotites et al., 1990; Johnson & Dick, 1992) exhibit systematic are apparently metamorphosed and deformed under variations with whole-rock MgO content. I also showed subsolidus conditions (e.g. Niu, 1997; Niu et al., 1997) that the whole-rock MgO content in melting residues is this could have resulted in mineral modal and comlinearly related to the overall extent of melting or melt positional re-equilibration. In this case, equation (1) could depletion [see fig. 8 of Niu (1997)]. These observations be an artifact of such a post-melting equilibration. The thus allow an explicit examination of how residual mineral effect of this equilibration may exist, but it is not petproportions change in response to varying extents of rographically evident, and it cannot explain the observed melting as shown in Fig. 1 [fig. 12 of Niu (1997)], and modal changes with increasing extent of melting (see allow an insightful understanding of the chemical and Dick et al., 1984; Niu et al., 1997). As the abyssal peridotites mineralogical consequences of actual mantle melting prostudied are from many locations (not a single location), cesses beneath ocean ridges. Figure 1 can be expressed and define much of the data range on a global scale, I in terms of a simple mathematical form [equation (8) of thus interpreted equation (1), with b > a, as representing Niu (1997)]: the net effect of most likely complex polybaric (vs isobaric) melting processes beneath ocean ridges (Niu, 1997). This 0·466 cpx+ 0·652 opx+ 0·049 spl interpretation is supported by decompression melting model results in which b > a [see fig. 6 of Niu (1997)], = 1·000 melt+ 0·167 ol. (1) and by the polybaric phase equilibria analysis which shows that opx (vs cpx) melting is further assisted by the incongruent melting of opx ⇒ ol+ SiO2 with decreasing Equation (1) says that the abyssal peridotite data (vs models) pressure [see fig. 7 of Niu (1997)]. Importantly, this tell us that in order to produce 1·000 mass unit of a interpretation is consistent with, and required by, the basaltic melt, approximately 0·466 mass unit of cpx, 0·652 fact that melting can occur beneath ocean ridges only mass unit of opx, and 0·049 mass unit of spinel are required because of fertile mantle upwelling and decompression (i.e. to melt whereas about 0·167 mass unit of olivine is required polybaric vs isobaric) in response to plate separation. to crystallize. Equation (1) is identical in form to isobaric My observation [Fig. 1 and equation (1)] and inmelting reactions determined experimentally (e.g. Kinzler terpretations have received aggressive criticisms from & Grove, 1992a; Baker & Stolper, 1994): a cpx+ b opx+ c Walter (1999) because my observation differs from isobaric spl = d ol+ 1 Melt, suggesting that mantle melting is melting experiments and his model melting reactions [see indeed incongruent. What differs between equation (1) and the isobaric melting reactions is the relative ‘melting his equations (2)–(5), table 1 and fig. 2]. In this reply, I