Morris et al. correctly point out that we misquoted their work and attributed to them observations that they did not report in Ramanaidou et al. (2003). Eager to acknowledge nearly two decades of descriptive work on Yandi-type orebodies by the Commonwealth Scientific and Industrial Research Organization, we inadvertently attributed both their and our textural observations to them. We apologize for that mistake. As for the petrographic descriptions, minerals and textures identified, and implications for the evolution of the Yandi deposit, we stand by our data and interpretations.
Morris et al. state that the terms “vitreous,” “botryoidal,” and “late-stage” cement do not accurately describe typical ore textures at Yandi or other channel iron deposits (CIDs). They also suggest that the material that we describe does not constitute typical Yandi ore.
The samples used for geochronology were collected from active mine faces at the Yandi deposit. Most samples were collected at the middle of the Yandi paleochannel, with some samples collected from the channel margins, as shown in Heim et al. (2006), Figure 1. Because it is a poly mictic marginal conglomerate, the sample shown in Data Repository Figure A1b (Heim et al., 2006) is not typical Yandi ore. However, it is a CID-facies and thus genetically linked to CID formation.
The Yandi CID samples contain fragments of detrital hematite, goethite, and ferruginized wood; some samples also contain ferruginized clay pods. All detrital grains are cemented together by late-stage goethite. Goethite cement varies from microns to centimeters in thickness, invariably shows a finely laminated texture, is highly crystalline and indurated, varies from dark brown to black on freshly broken surfaces, and has a vitreous lustre (DR Appendix 1, Heim et al., 2006). Concentric goethite cement surrounding clasts could be described as having a “colloform” texture. Clasts surrounded by colloform goethite and cemented by colloform late-stage goethite form a mass that resembles a “bunch of grapes;” therefore, “botryoidal texture” accurately describes this material. We interpret the goethite cement to be “late-stage” based on paragenetic relationships. Vitreous goethite binds together clastic grains and partially or completely fills pores (DR Appendix 1, Heim et al., 2006). When more than one generation of vitreous goethite exists, cross-cutting relationships are clear.
Therefore, we maintain that (1) Yandi-type ore contains vitreous goethite, (2) vitreous goethite is a late-stage cement, and (3) samples composed of detrital grains surrounded by vitreous goethite display a botryoidal texture.
Morris et al. suggest that we mixed original matrix and cortex with our late-stage goethite and doubt our ability to recover pure goethite from 1.5-mm growth zones with a 4-mm-diameter drill core. As described by Heim et al. (2006), the 4-mm drill core is the first stage of a systematic sample recovery and characterization process. After drilling, the 4-mm core is crushed to 0.1–3 mm grain size and sieved, washed in ethanol, and only pure goethite grains (easily recognized with the aid of a binocular microscope) are picked. We select 5–10 grains for geochronology, while an aliquot of visually pure grains are mounted in epoxy disks, described petrographically, and investigated under a SEM and an EM. Bench-top and synchrotron-based XRD of goethite cement extracted by this procedure confirms pure goethite concentrates.
We cannot completely reject the suggestion made by Morris et al. that some of goethite grains could be partially mixed with goethite from the cortices of pisoliths. However, even if small amounts of cortices were included, none of the samples that we have analyzed so far contain significant amounts of this detrital goethite. We stringently excluded texturally distinctive cortex phases from the picked sample aliquots, so any conceivable contamination must be very small. The poorest (U-Th)/He age reproducibility is observed for a sample with a 1-cm-thick late-stage goethite vein, which is petrographically devoid of any detrital phase. Therefore, the small but statistically significant age irreproducibility cannot be explained by admixture of pisolith cortices and warrants further investigation.
Morris et al. also accuse us of statistically forcing our data. There is very little statistical treatment of the data. We used linear regressions simply to illustrate how the (U-Th)/He age reproducibility for duplicate samples affects an extrapolation of the (U-Th)/He dates to the projected original surface of the channel. As stated in Heim et al. (2006), duplicate aliquots yield a narrow range of ages, but the uncorrected 4He ages are not within analytical uncertainty. This irreproducibility could have various possible sources: (1) we underestimated errors in 4He extraction and measurement, (2) we underestimated errors in U and Th analysis, (3) the samples lost various amounts of U and/or Th, (4) the samples lost different amounts of 4He during their geologic history or during sample preparation, and (5) the finely laminated vitreous goethite samples contain various generations of supergene goethite, spanning a range in ages. Procedures and the statistics for the treatment of analytical error in U and Th analysis by ICP-MS, and 4He analysis by mass spectrometry, are well-established and treated in Farley (2002). As stated in our paper, we cannot and do not address possible U and Th gains or losses. Some grains may indeed contain various proportions of supergene goethite from different generations; we can only address this issue by increasing spatial resolution. On the other hand, we can quantify 4He loss through 4He/3He geochronometry.
4He/3He experiments routinely show that natural goethites have diffusively lost 0% to a maximum of 20% of radiogenic 4He since precipitation. We applied the worst-case scenario to all samples. Extrapolating the geochronological data to the original channel surface illustrates an approach to estimate the end of aggradation and the onset of post-depositional goethite cementation of the channel sediments; we can only estimate this age within ± 20% margin of error.
Since the goethite cements that we investigated are demonstrably late-stage phases, they also provide clear minimum ages for the host sediments. The downward decrease in goethite cement ages evident in our data is not explained by any conceivable sediment depositional mechanism, as proposed by Morris et al. The geochronological results strongly suggest a pre-Miocene age for the channel sediments, which were subsequently altered by post-deposition weathering and goethite cementation to form CID ore at Yandi.