Nova Terra 52

of continental origin (Rojo-Pérez et al . 2022). However, the Hf model ages are not greatly older ( c . 1000 Ma), suggesting only a limited contribution of very old Gondwanan isotopic sources in the generation of these peri-Gondwanan magmatic rocks, either because of their lack in the source regions or because of their dilution by abundant juvenile material. 5.4. Garnet U-Pb geochronology U-Pb analyses were obtained on 34 points within a garnet crystal of sample 117,350 (Supplementary Table 5). Four analyses with a higher deviation were discarded to obtain a better fit (MSWD = 1.09), although the age difference using all the data is small. The lower intercept on the Tera-Wasserburg diagram provided an age of 590 ± 7.8/11.5 Ma (Figure 12a). A 238 U- 232 Th diagram shows a good cor- relation between both values (Figure 12b), indicating that all the analysed points correspond to a same mineral and that it is possible to rule out the presence of zircon microinclusions within the garnet, which can also be completely ruled out based on petrographic characteristics. Hence, the age of c . 590 Ma is consid- ered significant for garnet crystallization and therefore also for the age of the prograde metamorphic event affecting the Merida Ophiolite, the main tectonother- mal event that can be observed in this unit. 5.5. Mineral chemistry The garnets show a wide central area mainly devoid of zoning, which is restricted to a more or less narrow section, depending on the elements, near the rim of the crystal (Figures 4b to 4d). In the section near the rim, the most pronounced decrease in X Alm and X Sps content (core-rim: 0.65–0.57 and 0.05–0.02, respectively) can be observed, offset by an increase in X Grs content (core-rim: 0.12–0.21), the X Prp content being relatively constant, which generates an increase in the Mg/Mg+Fe ratio towards the rim. These patterns seem constant, since they are identical in the two profiles studied in different garnets (Figures 13a and 13b). It is difficult at first to interpret the nature of this zonation based exclu- sively on composition data, since it could be considered due in part to growth and in part to diffusional zoning (Tracy 1982). However, for a general interpretation in terms of reaction modelling, this zonation pattern can be interpreted as growth zoning generated during a prograde metamorphic regime (see below). However, all garnet composition project in the field of garnets characteristic of C-type eclogites, which may be an indication of a rather moderate crystallization tempera- ture (Figure 13c). Based on the classification of Hawthorne et al . (2012), the analysed amphiboles have pargasite compositions, with a few crystals showing compositions of Mg- hornblende or transitional to this group (Figure 13e) (Supplementary Table 7). The plagioclases are fairly homogeneous since all the analysed compositions cor- respond to andesine (Figure 13d). However, a very small compositional variation occurs between the plagioclase crystals included in the garnets (X An = 0.42–0.47) and those that appear in the matrix (X An = 0.45–0.48) (Supplementary Table 8). The analysed clinozoisites have variable pistacite contents between X Ps = 0.17 and 0.25 (Supplementary Table 9), with apparently somewhat higher contents in the crystals included in garnet in relation to those analysed in the matrix. 5.6. Pseudosection modelling The calculated pseudosection shows good consistency with the mineralogy of the garnet amphibolites. The region with the highest compatibility is delimited by the destabilization of Rt, Chl, and Ol and at temperatures below melting (Figure 14). Additionally, the absence of Opx and Cpx and the appearance of Qz and a single amphibole with the composition of pargasite, instead of the pair Prg + bcam (Prg + grunerite end member; Diener et al . 2007), constraints the stability of the amphi- bolites to the field marked with a star in Figure 14. In addition to phase assemblage equilibria, isopleths for X Grs , X Alm , X Sps , and X Prp in garnet (Supplementary Figures 2a to 2d) and X Ab in plagioclase (Supplementary Figure 2e) were calculated, together with the isopleths for the modal content of garnet in the amphibolite (Supplementary Figure 2 f). The calcu- lated isopleths show that the stability domains obtained for the end-member compositions of the garnet cores are quite scattered, which seems to indicate thermody- namic imbalance, although the core-rim variation of these compositions suggests progradation (Supplementary Figures 2a to 2d). Given these argu- ments, the core composition of garnets has not been used to deduce the P-T conditions of the metamorphic peak. Less dispersion is obtained with the compositions of the garnet rims and the average compositions of the plagioclases of the matrix, although the X Prp content is the one that differs the most (Supplementary Figure 2d). The projection of these values in the pseudosection, also taking into account the patterns of modal increase of the garnet content, defines a prograde P-T path that pro- gresses through a P-T region characteristic of medium pressures (Figure 15). It is unclear whether or not this INTERNATIONAL GEOLOGY REVIEW 19 Mérida Ophiolite (SW Iberia)