Stoup cut into the top of a Roman altar, the body of which has also been trimmed back to form a tapering shaft. The base is cut from a separate stone with different geology (see above) and the bottom has been trimmed at an angle to fit onto the sloping ledge on which it stands. The bowl of the stoup is rectangular with a sub-rectangular base and no drain hole. The sides and top of the upper part of the stone are decorated with slightly rounded parallel mouldings 4–4.5 cm (1.6–1.8 in) wide. There are four mouldings on each side face and two across each end of the top face. These are probably part of the decorative scheme of the original altar.

The upper part of the front face of the stone carries an inscription:

Other readings and translations have also been suggested (Collingwood and Wright 1995, cat. 304)

Appendix K item (Fonts and stoups in the Western Midlands).

This Roman altar has been significantly reworked to form what was almost certainly a pedestal-stemmed holy water stoup. The area occupied by the inscription seems, however, to have been quite carefully preserved, perhaps indicating that the lettering was thought to be important. This could be the result of a desire to maintain some form of symbolic link with the Roman past, or perhaps, if the local priest was literate, an assumption that the start of the inscription DEO TRI (the rest of the top line is damaged) might actually be a reference to the Trinity (for a fuller discussion of reused Roman material see Shrewsbury Abbey 1, p. 389). Michaelchurch is isolated deep in the Herefordshire countryside, more than sixteen miles from Kenchester, the nearest significant Roman town in the area (Magnis), but it is relevant to point to the tiny bowl cut into the top of a reused Roman column that currently serves the church there as a font but might also have originally been a holy water stoup (Kenchester 1, p. 382, Ills. 735–6). See also Chapter V, Further thoughts on fonts (pp. 62–4).

[1] There are, beside the Deerhurst font in Gloucestershire which has been shown to be of ninth-century date (Deerhurst St Mary 3, p. 163, Ills. 132–44, 740), a number of fonts in the study area that have been said to be Anglo-Saxon or could be Anglo-Saxon. There are also objects like Bisley All Saints 6 (below, Ills. 732–4) that has been described as a font fragment, and Kenchester 1 (p. 382, Ills. 735–6) that now functions as a font, but that are much smaller than all of the other vessels and may, therefore, have originally been used as stoups or lavabo bowls (see below, and 'Further thoughts on fonts' in Chapter V, pp. 62–4, Table 1). In the following Appendix three vessels that were probably stoups have been listed first, followed by the fonts in chronological order by form (cylindrical tub fonts, square tub fonts, tapering or cone-shaped fonts, and bowl-shaped fonts). Some clearly belong to the Overlap period but are included because they show continuity of form and decoration into the later decades of the eleventh century and beyond.

The tub font at Deerhurst is the earliest securely datable font, and an eighth-century Anglo-Saxon ivory panel in the Victoria and Albert Museum that depicts the baptism of Christ also show a tub font (Beckwith 1972, 119, cat. 5, ill. 20). Tub fonts have, therefore, been placed first in the catalogue below. However, in the south-west of England the earliest surviving fonts are bowl-shaped (copies of domestic bowls) and it seems inherently likely that both tubs and bowls were in use at the same time (Cramp 2006, 38; Blair 2010).

Many of the western Midlands fonts seem to have been carved from newly worked stone, but several are carved into reused Roman capitals and bases. One of the reused Roman bases (at Woolstaston, Shropshire), almost certainly came from the Roman city of Viroconium (Wroxeter) but, unlike the similarly reused bases at Wroxeter St Andrew and Shrewsbury Abbey (pp. 390, 389, Ills. 762–3, 768–70), this vessel has been very crudely reshaped and the bowl is only 8 cm deep (p. 386, Ills. 756–7). It does not look like a font at all but it would, in fact, be ideal for the baptism of adults by affusion or aspersion. Adult baptism must have been very rare by the later Anglo-Saxon period, so it seems possible that the Woolstaston font might be very early, perhaps even sub-Roman.