The Cretaceous Mount Daniel Complex (MDC) in northern Fiordland, New Zealand was emplaced as a 50 m-thick dyke and sheet complex into an active shear zone at the base of a Cordilleran magmatic arc. It was emplaced below the 20–25 km-thick, 125.3 ± 1.3 Ma old Western Fiordland Orthogneiss (WFO) and is characterized by metre-scale sheets of sodic, low and high Sr/Y diorites and granites. 119.3 ± 1.2 Ma old, pre-MDC lattice dykes and 117.4 ± 3.1 Ma late-MDC lattice dykes constrain the age of the MDC itself. Most dykes were isoclinally folded as they intruded, but crystallised within this deep-crustal, magma-transfer zone as the terrain cooled and was buried from 25 to 50 km (9–14 kbar), based on published P-T estimated from the surrounding country rocks. Zircon grains formed under these magmatic/granulite facies metamorphic conditions were initially characterized by conservatively assigning zircons with oscillatory zoning as igneous and featureless rims as metamorphic, representing 54% of the analysed grains. Further petrological assignment involved additional parameters such as age, morphology, Th/U ratios, REE patterns and Ti-in-zircon temperature estimates. Using this integrative approach, assignment of analysed grains to metamorphic or igneous groupings improved to 98%. A striking feature of the MDC is that only ~ 2% of all igneous zircon grains reflect emplacement, so that the zircon cargo was almost entirely inherited, even in dioritic magmas. Metamorphic zircons of MDC show a cooler temperature range of 740–640 °C, reflects the moderate ambient temperature of the lower crust during MDC emplacement. The MDC also provides a cautionary tale: in the absence of robust field and microstructural relations, the igneous-zoned zircon population at 122.1 ± 1.3 Ma, derived mostly from inherited zircons of the WFO, would be meaningless in terms of actual magmatic emplacement age of MDC, where the latter is further obscured by younger (ca. 114 Ma) metamorphic overgrowths. Thus, our integrative approach provides the opportunity to discriminate between igneous and metamorphic zircon within deep-crustal complexes. Also, without the tight field relations at Mt Daniel, the scatter beyond a statistically coherent group might be ascribed to the presence of “antecrysts”, but it is clear that the WFO solidified before the MDC was emplaced, and these older “igneous” grains are inherited. The bimodal age range of inherited igneous grains, dominated by ~ 125 Ma and 350–320 Ma age clusters, indicate that the adjacent WFO and a Carboniferous metaigneous basement were the main sources of the MDC magmas. Mafic lenses, stretched and highly attenuated into wisps within the MDC and dominated by ~ 124 Ma inherited zircons, are considered to be entrained restitic material from the WFO. A comparison with lower- and upper-crustal, high Sr/Y metaluminous granites elsewhere in Fiordland shows that zircon inheritance is common in the deep crust, near the source region, but generally much less so in coeval, shallow magma chambers (plutons). This is consistent with previous modelling on rapid zircon dissolution rates and high Zr saturation concentrations in metaluminous magmas. Accordingly, unless unusual circumstances exist, such as MDC preservation in the deep crust, low temperatures of magma generation, or rapid emplacement and crystallization at higher structural levels, information on zircon inheritance in upper crustal, Cordilleran plutons is lost during zircon dissolution, along with information on the age, nature and variety of the source material. The observation that dioritic magmas can form at these low temperatures (< 750 °C) also suggests that the petrogenesis of mafic rocks in the arc root might need to be re-assessed.