More recently, I have been able to examine a second sample – see Figs 8 to 10. Unfortunately, it had serious surface damage as a result of mishandling.  However, microscopic examination revealed some unique characteristics of H-glaze.

The Second Sample

The base material was a weakly compacted piece of chalk with a maximum dimension of about 3.5cm.   Three surfaces of this sample were covered with a thin deposit of brown colouration.    The deposits on two of these surfaces, Figs 9 & 10, were severely damaged through abrasion.   There was no deposit on the back of the sample.

The deposit on the top (largest) surface (Fig 8) had also suffered abrasion; however, it could be seen to be of moderately uniform thickness and to have none of the calciferous or silicone inclusions found in the first sample.  It had the appearance of a finely powdered precipitate but was cohesive and adhered well to the substrate. This glaze seemed to be composed largely of hydrated ferric oxide but, unlike rust, appeared not to have the complex structure of oxides and hydroxides found in corrosion product. Instead, it appeared to have a more refined composition.

Close examination revealed an intriguing feature.   In some areas, where the oxide coating was missing, there was a distribution of particles of elemental iron adhering tightly to the substrate.   These had the appearance of small individual solidified droplets and, where visible, they were dispersed uniformly over the surface of the chalk.  Most significantly, they were not corroded but had a bright metallic appearance.   They had plainly not contributed to the oxide that, presumably, had once covered them, as they showed no signs of corrosion.  If, as some suggest, the superficial layer had derived from naturally corroding particles of iron, the remaining elemental iron would have been corroded and tightly enmeshed in the corrosion product producing a single heterogeneous layer.  It would not have been possible to reveal clean bright metallic iron beneath it.
More interestingly, however, was the distribution of the iron, which was uniform and had the appearance of small separate individual solidified droplets.   The distribution and appearance of these elemental particles was highly suggestive of condensation from a vapour phase.   This, in turn, would have required temperatures well in excess of the melting point of iron.

With such samples as this, one would expect to be able to derive information from ‘shadowing’ and ‘splatter’ patterns in the deposit.   Unfortunately, the damage on two surfaces (Figs 9 & 10) make such observations highly speculative.   However, it does appear that the deposit had coated at least three mutually orthogonal surfaces of the sample.  Despite the damage, sufficient remained to see that there was no tapering of the glaze at the edges of the coated surfaces, which would have been the case if the deposit had been sprayed or dropped onto the chalk.