Kilpeck church is a three-cell building consisting of nave, chancel and rib-vaulted apse, all dating from the 12thc. At the west end of the nave is a wooden gallery including material dating from the 16thc. to the 19thc. There is no record to show when it was erected. The church is constructed of old red sandstone blocks; irregular in size and shape in the nave and chancel but of regularly coursed squared ashlar in the apse, which was refurbished by Cottingham in 1846 at the same time as its roofline was lowered. The walls are supported by flat, slender pilaster buttresses of ashlar, dividing the nave into three bays, the chancel into two, and the apse into one straight bay and three curved ones. There is no pilaster buttress at the NE angle of the nave, although traces of one remain at the top of the wall. Below, the angle is of long and short quoins and the nave wall to the N of the angle is of masonry different to that of the rest of the nave wall, sloping inwards so that it dies into the wall just below the level of the original nave window sills. This section of wall is on a slightly different line to the rest of the nave wall. Its interpretation, as a vestige of an earlier church on the site or as a later rebuilding, is discussed in section VIII. Over the west gable of the nave is a gabled double bell-cote; also part of the 1846 restoration.

Kilpeck is generally considered the jewel of the Herefordshire school of Romanesque sculpture; for its completeness, its virtuosity and its remarkable state of preservation. The south nave doorway, with a Tree of Life tympanum and richly carved jambs and arch orders, was protected by a wooden porch of unknown date until 1868. It has never been replaced, but a lead mantle was installed around the label in 1962 to prevent water penetration. The great west window is also elaborately carved. Of the other windows, those in bay 2 of the north nave wall, bay 1 of the south nave wall and the three curved bays of the apse are original. The south window of the chancel is 13thc., but two 12thc. corbels have been reused as label stops (described in section III.3.c.vii). A corbel table runs around the entire church, originally consisting of 91 corbels, most of which survive in excellent condition. They depict human and animal heads, birds, beasts and obscene subjects, some in a simple, almost cartoon-like style, others with classicising features, and all very easy to read since the eaves are not high. In addition there are projecting dragons' heads at the tops of the buttresses at the NW, SW and SE angles of the nave and in the centre of the west facade, all at the level of the corbel table.

Inside, the chancel arch has jamb-figures and chevron- and lozenge-decorated arch orders. The apse arch is plain, but the apse vault ribs are chevron-decorated, the vault boss is carved with lion heads, and there is sculpture on the inside of the apse windows. The font has an enormous plain bowl of conglomerate, and the church also contains a holy water stoup, imported from elsewhere, with a carved bowl and base; and a rare font-stopper carved with basketweave.

The monochrome photographs of corbels were taken on a Courtauld Institute of Art photographic trip in 1970; the colour photography was carried out in March 2005.

The earliest notice of a church at Kilpeck is in the Book of Llandaff, which records that the church and its lands were given to that diocese c.650. Kilpeck remained part of the Llandaff diocese until the 1130s, when it was appropriated by Hereford. The manor was held by the Conqueror's kinsman William fitzNorman in 1086, and no church or priest was noted at that time. William's son Hugh de Kilpeck is usually credited with the building of the church. In 1134 a small Benedictine priory founded there was given to St Peter's, Gloucester, and the grant refers to the castle, held by Hugh, its chapel of St Mary and the church of St David. The lordship passed to Hugh's son, Henry and thence to his grandson, John. King John visited the castle in 1211, 1212 and 1214. By 1259 it was held by Robert de Walerand, who was granted two annual fairs and a Friday market by Henry III. These grants were renewed by Edward III in 1309, when the lord was Alan de Plogenet.

Benefice of Ewyas Harold with Dulas, Kenderchurch, Abbeydore, Bacton, Kentchurch, Llangua, Rowlstone, Llancillo, Walterstone, Kilpeck, St Devereux and Wormbridge.

Kilpeck church has attracted several important publications of which the earliest are by G. R. Lewis (1840 and 1842). An earlier notice by Wathen (1789) is interesting in recording that the church was in a state of 'reparation and improvement' at that time. Lewis's detailed drawings of corbels have already proved useful in identifying features since lost (see corbels A3 and A8 above) and in identifying changes resulting from the restorations of 1846 and 1898 (e.g. the form of the apse roof). His work has sometimes been treated merely as a useful source for the appearance of the church in 1838, when the drawings were made, but this can be misleading because he had another agenda. In many of his general views, Lewis attempted to reconstruct the 12thc. appearance of the church in order to contrast its original 'fair and intelligent form' with its then current state as a 'mangled piece of confusion'. Hence, in his view from the S (pl. 15), he omitted the later chancel doorway and windows, and above all the wooden porch, 'it being in its whole appearance a public-house porch', and the W bell-cote, 'as its form is that of a beer-house chimney'. These features, now gone, are fortunately shown in pl.16.

Lewis's purpose in all this was an appeal for a restoration, which duly took place a few years later. The only general view he made of the N side (pl.8) was an attempt at reconstructing the original form, and he shows the nave divided into three bays by pilaster buttresses, including one at the NE angle which may never have existed. As noted in section II above, the angle has long and short quoins. The nave wall immediately alongside it has no plinth course, and it runs at a slightly different angle from the rest of the nave wall, so that where it ends there is a step back to the wall surface that dies away around windowsill level. RCHME (1931) described this as pre-Conquest work; the remains of an earlier church on the site, an interpretation accepted by Zarnecki (1950). This is the simplest explanation of the anomaly, and the one to which the present author inclines, but the situation is complicated by the fact that the masonry to the east of the angle, including the north wall of the chancel, has been subjected to later rebuilding work. An inserted window in the chancel wall is surrounded by later masonry, and the corbels above (NC1 to NC9) are by a different hand from any other corbels in the building. The corbel table itself is of the standard design, however. Both King (1992) and Thurlby (1999) argued that these features represent a post-medieval attempt to shore up a wall that was unstable, rather than the remains of Anglo-Saxon fabric. The Domesday Survey recorded no church at Kilpeck in 1086, but this is by no means conclusive evidence that there was no church there at that time.

The dating of the present building is contentious, and the historical record is of little value. It is certain that there was a church on the site in 1134, when Kilpeck Priory was founded as a daughter of St Peter's, Gloucester; the grant referring to the castle, its chapel of St Mary and the church of St David. Both King and Thurlby have argued that St David's was the present church on the grounds that no earlier one existed. King further notes that a dedication to St David would be unlikely in England before his canonisation in 1120. Thurlby also argues that the Welsh rising of 1136 and the civil war that followed the death of Henry I in 1135 make it unlikely that a new church would be begun in the years immediately after that date. King's lengthy analysis leaves the reader doubting whether any building work at all could have taken place in the West Country between 1138 and 1147, in view of the conflicts recorded in that period. Such hypotheses are impossible to prove or to disprove in the absence of precise dates for the churches of the Herefordshire School. We come closest to this in the case of Shobdon, consecrated when complete by Robert de Bethune, Bishop of Hereford (1131-48), probably before 1143 when Robert quarrelled with Miles of Hereford and was obliged to take refuge in Shobdon. Opinions about the date of Kilpeck are now polarised, and largely depend on the view taken of the development of the Herefordshire workshops. Zarnecki's more recent views place the church c.1140. King places Kilpeck earlier than Shobdon, offering a date in or before 1134. Thurlby agrees that the church was complete by 1134, but suggests that Shobdon was a few years earlier. Pevsner dates the start of the campaign shortly after 1134 but suggests that the sculptors may have moved to Shobdon c.1140 and returned to Kilpeck c.1145 when Shobdon was finished.

The first detailed analysis of the sculpture was Zarnecki (1950), still a key work for the Herefordshire School, and still unpublished. A more detailed description of the school will be found in the Preface to Herefordshire. Zarnecki identified the work of two main carvers; one who had worked alone at Aston and Ribbesford (Worcs), he christened the Aston Master; the other, first seen at Shobdon working alongside the Aston Master in a bold and flamboyant style, he called the Chief Master. The Chief Master's work is exemplified by the apostle figures on the jambs of the chancel arch at Kilpeck. These two main personalities were accepted by Thurlby, and he and Zarnecki postulated a long-term relationship of collaboration between the two, and the presence of other, more minor figures. Zarnecki identified common motifs at Kilpeck and Shobdon, for example the serpents on the jambs of the S doorway, seen on the left doorway label and on shaft 5 at Shobdon, and the warriors on the S doorway found on shafts 2 and 8 at Shobdon. The column-figures on the chancel arch, with their bulging eyes and thin moustaches, he compared to climbing figures on shafts 2 and 8 at Shobdon, attributing them to the Chief Master. The drapery of these figures is more plastic than the simple grooving of the warriors on the doorway, but Zarnecki again found a Shobdon comparison (shaft 3). He also suggested a manuscript comparison (Hereford Cathedral Library P.I.7). This probably post-dates Kilpeck, and may thus have been copied from there. It is an important feature of Zarnecki's analysis that at least one of the Shobdon sculptors accompanied Oliver de Merlimond on a pilgrimage to Santiago de Compostela, and that Shobdon, for Zarnecki the earliest church of the Herefordshire School, was completed on his return, incorporating ideas picked up on a journey that would have reached Santiago via western France. The disposition of the chancel arch figures, one above the other, he attributed to the sculptor having seen the Puerta de las Platerias in Santiago de Compostella. He attributed the label of the S doorway to the Aston Master, and the birds carved there are similar to those on Shobdon shaft 6, while the first three lions' heads closely resemble those on Shobdon font (where they are attached to lions). Birds similar to those on the label at Kilpeck also appear at Brinsop, Rowlstone, Leominster and on the Castle Frome font. The beakhead arch of the S doorway is unique for the Herefordshire School, and this Zarnecki attributed to a sculptor from Reading Abbey. Grander claims for the importance of Reading to the school are found in Jonsdottir (1950), but the comparisons made there fail to demonstrate that Reading had any influence on the more distinctive features of Herefordshire sculpture. The decoration of the S doorway label with motifs in rings is rare in England, Zarnecki noted, but commoner in western France, e.g. on the arch of the W window of Notre-Dame-la-Grande at Poitiers.Foliage tympana like Kilpeck's are not found elsewhere in the work of the Herefordshire School, but are among the distinctive features of the Bromyard (or Dymock) group, as at Yatton and Dymock. Other features inherited from that group are the type of capital found on the S embrasure of the chancel arch, combining scallops with a ring of foliage (cf. Dymock) and the use of a plain inner order to the doorway (cf. Yatton and Moccas).

Kilpeck has often been seen as an important manifestation of Scandinavian ideas in England, but for Zarnecki they appeared only in the curious form of the outer-order jambs of the doorway, found in timber doorways in Norway, with flat jambs shafted on their inner sides, and in the dragons' head projections around the nave (cf. Borgund, Gol). The only other English example of a beast projecting in this way is at Barton Seagrave (Northants), which is certainly older. He was not convinced by comparisons made between the dragons on the outer jambs at Kilpeck and Urnes-style work, because the resemblances are not close, and most of the Scandinavian examples are later 12thc.

The corbel table of Kilpeck is the only one surviving in-situ carved by the Herefordshire School. Zarnecki recognised the hands of both main sculptors, and pointed out the presence of a sculptor working in a very classicising style (see e.g. NN1-2, NW11-12), which he compared with a Romano-British head in Gloucester Museum. In general he found a greater uniformity of style at Kilpeck than at Shobdon, which he attributed to the fact that the sculptors had been collaborating for longer and their artistic difference had grown less marked. The western French comparisons seen in profusion at Shobdon were reduced here, although the borrowing from Santiago itself was innovative. There were more local links, e.g. with the Bromyard group, and some influence from Scandinavian timberwork.

Responses to Zarnecki's masterly analysis have been concerned, it seems, either to question his chronology or to evaluate his attribution of the sources of the multifarious output of the Herefordshire sculptors, although he was always concerned to emphasise their inventiveness; never suggesting that they were either copyists or synthesisers.

King was not convinced of the close connection between Kilpeck and Shobdon, preferring to emphasise Kilpeck's uniqueness within the school and finding the closest local comparisons with the weathered tympanum of St Giles's, Hereford. He rejected all notions of Scandinavian influence, arguing that the dragons' head projections could be just as easily explained in terms of Anglo-Saxon work, exemplified by the heads on the west front of Deerhurst (Glos). He was similarly concerned to emphasise that a good deal of what was seen to be distinctive about Kilpeck was already part of the stock-in-trade of English sculptors by the 1130s, pointing to Castor (Northants) with capitals bearing foliage-spewing heads on the angles, to Hereford Cathedral (backing his case with reference to cathedral manuscripts), to Llandaff Cathedral, and to Lullington (Somerset) which has a range of corbels paralleling Kilpeck's and a tympanum with a beaded-stemmed tree. Lullington has workshop links with Old Sarum, which itself provides a very close comparison for one of the Kilpeck apse corbels (A12). The unusual bead on the forehead of A12 is also found at Old Sarum. Old Sarum only survives as fragments, but was apparently the source for sculpture at Leonard Stanley (Glos) and Great Durnford (Wilts), which both provided King with further parallels with Kilpeck. King rejected Zarnecki's suggestion that Reading may have been important, preferring Old Sarum as the source of the beakhead ornament on the S doorway.

Like Zarnecki, King emphasised the importance of western French sculpture to Kilpeck, but his comparisons were with the churches of the Saintonge, e.g. St-Eutrope at Saintes, St-Jouin-des-Marnes and Thouars, where he identified the motif of the head spewing foliage from its mouth and ears. For him the most important churches in this region were the pilgrimage sites of St-Eutrope at Saintes and St Jean d'Angely. The latter has gone, but fragments remain that he associated with Kilpeck. For King the famous pilgrimage of Oliver de Merlimond is less important than the fact that there were direct ties between England and Saintonge as early as the 1120s. He put forward the example of Henry of St Jean d'Angely, a kinsman of Henry 1; abbot of St Jean d'Angely (1103/04-31), and simultaneously abbot of Peterborough (1127-32) as an example of another specific contact, to demonstrate that Oliver's case was by no an means isolated one.

Thurlby systematically tracked through the range of sources suggested for the School generally. He found many parallels with Kilpeck at Hereford Cathedral, including bases with beaked spurs, imposts with saltire crosses, cable necked capitals, stringcourse motifs and beaded medallions linked by masks (on a capital in the N nave arcade). He also convincingly related the chevron-decorated vault ribs in the apse at Kilpeck to ribs at Hereford (without finding precise comparisons). A curious feature of many of the head corbels at Kilpeck is the nasolabial ridge, and Thurlby found examples of this on capitals of the S transept at Hereford (Thurlby fig.10, cf corbel A14). A lost capital from Hereford recorded by the Royal Commission (RCHME (1931), pl.149) relates closely to the W capital of the S doorway at Kilpeck, suggesting to Thurlby that the Chief Master worked at Hereford Cathedral before going to Kilpeck. Another local source suggested by Thurlby is Gloucester Cathedral (more fully explored in Chwojko and Thurlby (1997)). Frontal chevron, seen on the Kilpeck S doorway and chancel arch appears in the N nave arcade at Gloucester. He also found figural and foliage comparisons at Kilpeck with Gloucester and Tewkesbury work which he dated to the early 1120s, before Kilpeck was given to Gloucester in 1134, and suggested that Gloucester-trained craftsmen were later employed at Kilpeck. He accepted and amplified Zarnecki's opinion on the importance of the Bromyard /Dymock School, and also argued for the importance of pre-Conquest and Romano-British sculpture. Thurlby accepted the importance of the Puerta de las Platerias at Santiago in the design of the superimposed figures on the chancel arch at Kilpeck, but was also prepared to countenance the possibility that work at Ferrara had its effect too, as suggested by Pevsner (1963). Similarly, for the beakheads at Kilpeck, Thurlby is prepared to accept that both Reading (as Zarnecki suggested) and Old Sarum (as suggested by King) were influential.

Kilpeck has attracted a good deal of iconographic speculation, particularly aimed at its corbels. Even the dragons' heads have been questioned. Druce (1909) argued that they were not dragons at all but crocodiles sleeping with their mouths open, and their spiralling tongues were the tails of their enemies the hydri that, according to accounts in Bestiaries, crawled into the crocodiles' bodies and ate their way out through the stomach, destroying their sleeping prey in the process. I see no evidence at all to support this bizarre assertion, so it is interesting to find it repeated as the truth in as recent a publication as Bailey (2000). The hydrus story is interpreted in the Bestiary text as symbolic of Christ's Harrowing of Hell, which explains the attraction of this interpretation to the writer of a popular church guide. The most famous of the corbels is, of course, the sheela-na-gig (A2), which occupies a prominent place in treatments of this subject such as Andersen (1977) and Weir and Jerman (1986). Both Bailey (2000) and the author of A Short Tour... were concerned to find thematic groups within the corbels. Bailey, for example, sees in corbels A18-A21 imagery of the fair (Kilpeck was granted a fair in 1259, but may have held one earlier) in the form of music, dancing, wrestling and catching a greasy pig. The best that can be said of this interpretation is that it is imaginative.

There is a high proportion of beast corbels, and Thurlby has attempted where possible to identify them, in order to provide them with their Bestiary moralisations. In some cases the identifications themselves can be disputed. Thurlby's bold interpretation of the upside-down beast head at A6 as an ibex (which carries a good deal of iconographic weight in Bestiaries) is surely refuted by the observation that it has antlers rather than horns and is likelier, as the author of A Short Tour... suggests, to represent a dead stag after the hunt. The author is reluctant to accept that Bestiaries provided the key to much medieval sculpture (Baxter 1998). They provided exemplars for sermons, which might well have been used to explicate such sculptures as these, but they were not the only source, and in any case their theological arguments are often so convoluted as to be almost incomprehensible. My own position is closer to that of the late Michael Camille, who argued that the corbels surrounding a church in a ring of imagery that was often worldly and sometimes grotesque and obscene could represent the reality of the world itself that surrounded and besieged the Church and the Man of God. Here at Kilpeck are images of everyday creatures, apparently pretty and appealing like the dog and rabbit (A7), and of pleasurable activities like music (A18), dancing (A21) and lovemaking (A21). These are juxtaposed with images intended to reveal the truth that all such pleasures are the snares of the devil: the gruesome sexuality of the sheela-na-gig (A2), and the devil himself who feeds on his prey, the soul of the sinner (A9). The only unambiguously religious corbels are A13 and SN8, both depicting the Agnus Dei and both in key positions in the table; opposite the altar on the axis of the apse, and above the south doorway - the only entrance available for the parish.

Of the other important sources, Hamer (1992) - an unpublished PhD dissertation from the University of Chicago - was unfortunately not available for consultation, but was used by Thurlby, who acknowledged its importance as a pioneering patron-based approach to the Herefordshire School.

The font, one of eight large bowls of breccia conglomerate in the county is dated c.1140-50 by Halsey, but Zarnecki prefers a date in the third quarter of the 12thc.