St Mary's, Deerhurst is certainly the most imposing, enigmatic and controversial Anglo-Saxon building in the county, and perhaps the whole country. It retains features from several periods of Anglo-Saxon building, covering the period (depending on which version of its history you prefer) from the 7thc to the 11thc. Putting the standing evidence together into a believable building sequence is by no means straightforward. The general reader who has prepared for a visit to the church might well be surprised to find any post-Conquest material there at all, but the interior is dominated by the late-12thc nave arcades that will be the main subject of this report.

An oversimplified version of the story is that the earliest church, predating the first mention of a monastery here in 804 might have been a rectangular box with a W porch. This church was greatly enlarged in a second phase of building, usually assumed to correspond to the late-10thc reform movement. Porticus were added alongside the nave, and a polygonal apse was added at the E end. The nave was raised at the same time to its present height of approximately 12m, and a chancel was carved out of the rectangular nave by the insertion of a cross wall. Around this time too, the W porch was given extra stioreys to convert it into a tower porch (although this may have been done in several phases). After the Conquest - perhaps as late as 1190 - the porticus flanking the nave were replaced by aisles with 3-bay arcades. The S aisle was later (but still in the 12thc) fitted with transverse arches carried on responds abutting the nave arcade piers. The plan included here, displayed inside the church, shows one interpretation of the building phases.

What survives, therefore, is a church with a chancel, aisled nave and a W tower porch. The aisles extend W alongside the tower, and a passage linking them, W of the central vessel, forms a kind of W transept. Remains of 12thc work survive here, notably in a section of foliate stringcourse of the W wall of the passage. The W end of the N aisle now houses Deerhurst's famous font; presented as a work of 9thc art in its own right, dramatically lit from below. This report is provisional, as access to the upper levels of the tower was not available at the time of the visit.

The earliest mention of the monastery of Deerhurst is in a chartulary compiled at Worcester in Wulfstan's time (1062-95), stating that there was a monastery there soon after 804. The house may have been burned by the Danes and rebuilt c. 970. In 1059-60, Edward the Confessor divided it between his new foundation of Westminster Abbey and Saint-Denis in Paris. It thus became an alien priory, and Prior Baldwin, who came from Saint-Denis, later became Abbot of Bury St Edmunds.

The arcade capitals provide a good example of the interplay between the traditional scallop form and the newer stiff- leaf. Trumpet scallops do not appear before c.1170, but at Deerhurst the sculptiors were sufficiently familiar with the form to treat it playfully; hyphenating the shields and using running foliage scrolls to define them. If we accept an evolutionary model of the form, it seems clear that the less adventurous forms of the N arcade preceded the flamboyant excesses of the S. We are privileged to see at close hand the way in which a Romanesque motif changes into a Gothic one, and the handling is so assured that no sense of rupture is distinguishable. The three male heads among the S arcade capitals; two monkish and the third hairy and fleshy (especially the pair on pier 2) certainly point out a contrast to the modern viewer, but it could be a mistake to project our own perceptions into the past.

As to the relative dating of the two aisles, Butterworth's earliest view (Butterworth (1862), was that the remodelling began in the mid-12thc, and that the two phases of work in the S aisle were evidence of the change from Transitional to First Pointed during a single campaign that included both arcades. Buckler (1887) also opted for a mid-12thc start date, but offered a more complex chronology The earliest work would be the transverse arch responds and springers behind the S arcade piers, along with a S arcade that was to be replaced later. There was a building break until the early 13thc, when the N arcade was built and a new S arcade to the same design was installed to replace the mid-century one. Butterworth revised his views in his publication of 1887, and his new analysis was much closer to Buckler's. More recent writers have adopted a simpler view. Verey (1970) dated the nave arcades c.1200 with the S a little later than the N, and Brooks's 2002 revision of Verey's work added that the responds attached to the S side of the S arcade piers could be the remains of the former porticus walls. The current view is handily summed up in Thurlby (2014, 31-46). He opts for a continuous single campaign, beginning with the responds on the S side of the S arcade piers, pointing out that the responds do not course with the piers except at the W side of pier 2. It follows that while the pier 1 S respond was constructed separately from the pier, and by the time the W side of pier 2 was reached the pier and its S respond were being built together. He also offfers a date range based on comparison with local examples. The multi-scalloped S capital of the S respond of pier S2 is similar to work at Ledbury (Herefordshire) dateable by comparison with Keynsham to 1166/67. This is backed up by comparisons with Little Malvern , Llanthony Prima and Worcester Cathedral to support a start date for Deerhurst around 1170-80.