Minor punctuation errors have been corrected without notice. An obvious printer error has been corrected, and it is listed at the end. All other inconsistencies are as in the original. The author's spelling has been maintained.
The Cambridge Manuals of Science and Literature
THE GROUND PLAN OF THE ENGLISH PARISH CHURCH
LONDON: FETTER LANE, E.C.
C. F. CLAY, MANAGER
EDINBURGH: 100, PRINCES STREET
BERLIN: A. ASHER AND CO.
LEIPZIG: F. A. BROCKHAUS
NEW YORK: G. P. PUTNAM'S SONS
BOMBAY AND CALCUTTA: MACMILLAN AND CO., LTD.
All rights reserved
THE GROUND PLAN OF THE ENGLISH PARISH CHURCH
A. HAMILTON THOMPSON M.A., F.S.A.
Cambridge: at the University Press 1911
PRINTED BY JOHN CLAY, M.A. AT THE UNIVERSITY PRESS
With the exception of the coat of arms at the foot, the design on the title page is a reproduction of one used by the earliest known Cambridge printer, John Siberch, 1521.
There is as yet no book entirely devoted to the development of the plan of the parish church in England, and the body of literature which bears upon the subject is not very accessible to the ordinary student. The present volume is an attempt to indicate the main lines on which that development proceeded. It is obvious that, from necessary considerations of space, much has been omitted. The elevation of the building, and the treatment of its decorative features, window-tracery, sculpture, etc., belong to another and wider branch of architectural study, in which the parish church pursues the same line of structural development as the cathedral or monastic church, and the architectural forms of the timber-roofed building follow the example set by the larger churches with their roofs of stone. To this side of the question much attention has been devoted, and of late years increasing emphasis has been laid on the importance of the vaulted construction of our greater churches, which is the very foundation of medieval architecture and the secret of its progress through its various "styles." It is expected that the reader of this book, in which a less familiar but none the less important topic is handled, will already have some acquaintance with the general progress of medieval architectural forms, with which the development of the ground plan keeps pace.
Some historical and architectural questions, which arise out of the consideration of the ground plan, and have an important bearing upon it, are treated in another volume of this series, which is intended to be complementary to the present one.
The writer is grateful to his wife, for the plans and sketches which she has drawn for him, and for much help: to Mr C. C. Hodges and Mr J. P. Gibson, for the permission to make use of their photographs; and to the Rev. J. C. Cox, LL.D., F.S.A., and the Rev. R. M. Serjeantson, M.A., F.S.A., for their kindness in reading through the proofs and supplying suggestions of the greatest value.
A. H. T.
26 January 1911
THE ORIGIN OF THE CHURCH PLAN IN ENGLAND
1. The basilican church plan 1
2. Problem of its derivation 2
3. Rival theories of its origin 3
4. The Roman basilica: old St Peter's 6
5. Basilicas at Ravenna 8
6. Tomb-churches and baptisteries 9
7. Centralised plans at Ravenna 10
8. Relative advantages of the basilican and the centralised plan 12
9. The basilican church at Silchester 13
10. Early churches in Kent and Essex 14
11. Bradford-on-Avon, Wilts. 16
12. Escomb church, Durham 16
13. Early Northumbrian churches 18
14. Wilfrid's churches at Hexham and Ripon 20
15, 16. Brixworth, Northants: other basilican plans 21
17. Exceptional occurrence of the basilican plan in England 24
PARISH CHURCHES OF THE LATER SAXON PERIOD
18. The normal pre-Conquest plan 27
19. The western bell-tower 29
20. Plans in which the ground floor of the tower forms the body of the church 30
21. Barton-on-Humber and the centralised plan 33
22. Centralised planning in England 34
23. The Saxon lateral porch 35
24. Development of the transeptal chapel 36
25. Towers between nave and chancel 37
26, 27. Development of the cruciform plan 38
28. Influence of local material upon the aisleless church plan 42
THE AISLELESS CHURCH OF THE NORMAN PERIOD
29. Survival and development of the aisleless plan after the Conquest 44
30. The nave of the aisleless church 46
31. Rectangular chancels 47
32. Churches with no structural division between nave and chancel 49
33. Churches with apsidal chancels 49
34. The quire 53
35. The transeptal chapel 54
36. Cruciform plans: North Newbald and Melbourne 58
37. Later developments of the cruciform plan 60
38. Symbolism in planning 62
THE AISLED PARISH CHURCH
I. NAVE, TOWER, AND PORCHES
39. Survival of the aisleless plan 64
40. The addition of aisles 66
41. Use of aisles for side altars 66
42. Twelfth century aisled plans 69
43. Ordinary method of adding aisles 70
44, 45. Consequent irregularities of plan 74
46. Gradual addition of aisles 77
47. Raunds church, Northants 79
48. Conservative feeling of the builders for old work 81
49. Aisles widened and rebuilt 83
50. Rebuilding of aisles as chantry chapels: Harringworth, Northants 84
51. Newark, Cirencester, Northleach, and Grantham 87
52. Naves lengthened westward 92
53. The western tower in relation to the plan 94
54. Engaged western towers, etc. 96
55. Rebuilding of towers 98
56. Porches 99
57. Position of the porch in the plan 99
THE AISLED PARISH CHURCH
II. TRANSEPTS AND CHANCEL
58. Cruciform churches with aisled transepts 101
59. Addition of transeptal chapels 102
60. Variety of treatment of transeptal chapels 105
61. Transeptal chapels as a key to original ground plans 107
62. Incomplete cruciform plans 108
63. Irregular cruciform plans 110
64. Central towers with transeptal chapels 113
65. Transeptal towers 113
66. Lengthening of chancels 114
67. Encroachment of the chancel on the nave: Tansor 115
68. Chancel chapels 117
69. Churches with one chancel chapel 119
70. Chantry chapels attached to chancels 120
71. Effect of the addition of chapels on the cruciform plan 121
72. The aisled rectangular plan 124
73. Variations of the plan with aisled nave and chancel 126
74. Development of the aisled rectangle at Grantham 129
75. Deviation of the axis of the chancel 131
INDEX OF PLACES 134
LIST OF ILLUSTRATIONS
Hedon. Interior of nave Frontispiece
1 Plan of old St Peter's 6
2 Plan of San Vitale, Ravenna 11
3 Plan of Escomb—typical Saxon church 17
4 St Peter's, Barton-on-Humber 31
5 Aisleless plan, 12th cent. 45
6 Birkin, Yorkshire: interior 51
7 Two aisleless plans with central tower 55
8 North Newbald 57
9 Sketch of older wall above nave arcade, Gretton 72
10 Plan of Raunds church 80
11 Plan of Harringworth church 85
12 Two plans, nos. 1 and 2, of Grantham church 88
13 Sketch of arch joining arcade to tower, Gretton 93
14 Plan of 13th cent. church: W. tower, S. Porch, transeptal chapels 103
15 St Mary's, Beverley. Interior of transept. 111
16 Plans of Grantham church, nos. 3 and 4 130
THE ORIGIN OF THE CHURCH PLAN IN ENGLAND
Sec. 1. Side by side with the establishment of Christianity as the religion of the Roman empire, there appeared a fully developed plan for places of Christian worship. The normal Christian church of the fourth century of our era was an aisled building with the entrance at one end, and a semi-circular projection known as the apse at the other. The body of the building, the nave with its aisles, was used by the congregation, the quire of singers occupying a space, enclosed within low walls, at the end nearest the apse. In the apse, raised above the level of the nave, was the altar, behind which, ranged round the wall, were the seats for the bishop and assistant clergy. This type of church, of which the aisled nave and the apse are the essential parts, is known as the basilica. The name, employed to designate a "royal" or magnificent building, had long been applied to large buildings, whether open to the sky or roofed, which were used, partly as commercial exchanges, partly as halls of justice. It is still often said that the Christian basilicas were merely adaptations of such buildings to sacred purposes. Some of the features of the Christian plan are akin to those of the secular basilica. The apse with its semi-circular range of seats and its altar reproduces the judicial tribune, with its seats for the praetor and his assistant judges, and its altar on which oaths were taken. The open galleries, which in some of the earliest Christian basilicas at Rome form an upper story to the aisles, recall the galleries above the colonnades which surrounded the central hall of some of the larger secular basilicas. Again, the atrium or forecourt through which the Christian basilica was often approached has been supposed to be derived from the forum in connexion with which the secular basilica was frequently built.
Sec. 2. However, while the atrium of the Christian basilica is merely an outer court, the secular basilica, when planned, like the Basilica Ulpia at Rome, with direct relation to a forum, was a principal building in connexion with the forum, but not a building of which the forum was a mere annexe. Further, when we begin to seek for a complete identification of the Christian with the secular basilica, we are met by the obstacle that the secular basilica had no fixed plan. If we try to trace any principle of development in its plan, we find that this development is directly inverse to that of the Christian basilica. The secular basilica, in earlier examples a colonnaded building with its central space open to the sky, became at a later time a roofed hall, either, as in the case of the basilica at Trier, without aisles, or, like the basilica of Maxentius or Constantine in the Roman forum, with a series of deep recesses at the side, the vaulted roofs of which served to counteract the outward pressure of the main vault. The Christian basilica, if it were a mere imitation of this type of building, would follow the same line of development; but, as a matter of fact, the highest type of Christian church is always a colonnaded or aisled building. And, even if the Christian apse derived its arrangement from the apse or apses which projected from the ends or sides of the secular basilicas, there is again a difference. The apse with its altar was the main feature of the interior of the Christian church: it was the place in which the chief rite of Christian worship was performed before the eyes of all. In the secular basilica the apse was devoted to special purposes which set it apart from the main business of the body of the building: it was an appendage to the central hall, not necessarily within view of every part of it. In fact, the relation of the apse to the main building was totally different in the two cases.
Sec. 3. It seems probable, then, that the identity between the two buildings is mainly an identity of name, and that Christian builders, in seeking for suitable arrangements for public worship, may have borrowed some details from the arrangements of the secular basilica. It is natural, however, to look for the origin of a religious plan in buildings devoted to religious purposes. The Roman temple supplied no help for the plan of buildings which were required for public worship. Of recent years, it has been customary to assume that the Christian basilica took its form from the inner halls of the private houses of those wealthy citizens who embraced Christianity in its early days. Such halls may have been used for Christian services; and if their plan was adopted for the Christian basilica, the mature state of the basilican plan at its first appearance can be explained. The atrium or entrance hall of the house is represented on this hypothesis by the forecourt of the basilica; the peristyle, or colonnade round the inner room, becomes the aisles and the space screened off at the entrance for those not entitled to take full part in the service; the colonnade at the further end survives in the arcaded screen which existed, for example, in old St Peter's at Rome; the apse takes the place of the tablinum, where the most sacred relics of family life were preserved; and the transept, which is found in some of the early Roman basilican plans, represents the alae, or transverse space, which existed between the tablinum and the main body of the hall. But these close analogies are the result of an assumption by no means certain. It is always probable that the basilican plan had its origin in a plan originally aisleless. Some, intent on its religious source, explain it as a development of the plan of the Jewish synagogue. Others, regarding assemblies of Christians for public worship as, in their essence, meetings of persons associated in common brotherhood, have derived the basilica directly from the aisleless scholae which were the meeting-places of the various confraternities or collegia of ancient Rome. In these there is an apse at one end of the building; and, if we imagine aisles added by the piercing of the walls with rows of arches and columns, we have at once the essential features of the basilican plan. Each theory has its attractions and its difficulties; and to none is it possible to give unqualified adherence. It may be stated, as a tentative conclusion, that the basilican plan probably had its origin in an aisleless form of building, and thus pursued a course directly opposite to the development of the secular basilica. But it seems clear that, in many details of the plan, especially as we see it in Rome, the peristyled hall was kept in mind; while in two features, the arrangement of the apse and the occasional appearance of galleries above the aisles, the secular basilica was taken into consideration. The policy of the early Christian Church, when its services were sanctioned by the state, was to adapt existing and familiar forms where they could be suitably reproduced.
Sec. 4. The plan of the old basilica of St Peter at Rome, founded by Constantine the Great, and destroyed early in the sixteenth century to make way for the present church, explains the principal features of the basilican plan in its developed state. (1) In common with other early basilicas in Rome, and in other parts of western Europe, the entrance was at the east, and the altar at the west end, so that the celebrant faced the congregation during the divine office. (2) The church was approached through a cloistered atrium or fore-court, in the middle of which was a fountain, the place of purification for those intending to enter the church. (3) At the west end of the cloister three doorways opened into the nave of the church, and one on either side into the side aisles. (4) The nave communicated with the aisles by a row of columns beneath an entablature: there were also outer aisles, communicating with the inner by columns bearing rounded arches. (5) The side walls of the nave, above the entablature, were not pierced for galleries, but were covered by two rows of mosaic pictures, one above the other, on each side, the upper row corresponding to the height of the space between the outer and inner roofs of the aisle. Above this, the walls rose into a clerestory, pierced with round-headed windows at regular intervals; and a high entablature supported the great tie-beams of the wooden roof. (6) The quire of singers, divided from the rest of the church by low screen walls, probably occupied the centre of the western portion of the nave. (7) A tall open arch divided the nave from the transept, which was of equal height with the nave, and projected south and north as far as the walls of the outer aisles. Here probably were places reserved for distinguished persons, near the platform of the altar. (8) West of the transept, entered by a tall and wide arch, was the apse. Beneath the arch was a screen, formed by a row of columns, under an entablature which bore statues of our Lord and the apostles: this crossed the arch at the foot of the steps leading to the altar and seats of the clergy. (9) Beneath the altar platform, and entered by doorways on each side of the flight of steps, was the crypt or confessio, the traditional place of martyrdom of St Peter, and the resort of pilgrims to the tomb of the apostles. The hallowed place was immediately beneath the altar.
Sec. 5. The sixth century basilicas of Ravenna, Sant' Apollinare in Classe and Sant' Apollinare Nuovo, differ in plan from the Roman basilicas (1) in the fact that they have always had the altar at the east, and the entrance at the west end; (2) by substituting, for a colonnaded atrium, a closed porch or narthex in front of the entrance of the building. In process of time, two of the greater Roman basilicas, San Paolo and San Lorenzo fuori le Mura, were enlarged in a westward direction, so that the positions of the altar and entrance were reversed; and, in several of the early basilicas at Rome, a space near the entrance of the nave was screened off, from which penitents and catechumens might watch the service. But, in the first instance, the eastern chancel and the structural narthex appear to have been introduced from the eastern empire. Neither at Ravenna nor at Rome did bell-towers originally form part of the plan of the basilica: the round campanili of both churches at Ravenna are certainly later additions. It may also be noted (1) that ordinarily the aisles were single, not double as at old St Peter's. (2) The columned screen of the apse at old St Peter's appears to have been exceptional. The ordinary screen or cancelli, from which is derived our word "chancel" for the space thus enclosed, was a low wall. This is the arrangement at the basilica of San Clemente, in which the enclosed quire also remains. (3) The transept, even in Rome, was an exceptional arrangement, and does not appear in the basilicas of Ravenna.
Sec. 6. Another type of plan, however, was used in Rome for churches devoted to the special purposes of burial and baptism. In this case the buildings were planned round a central point, and at Rome were uniformly circular. Recesses round the walls of the mausoleum-church contained sarcophagi: in the centre of the baptistery was the great font. The church of Santa Costanza, outside the north-eastern walls of Rome, circular in plan, with a vaulted aisle surrounding the central space, was built by Constantine the Great as a tomb-church for his family, and was also used as a baptistery. Both these uses were direct adaptations of pagan customs. The baptistery, with its central font for total immersion, was simply a large bath-room, like the great rotunda of the baths of Caracalla. The mausoleum preserved the form of which the finest example is the tomb of Hadrian, now known as the castle of Sant' Angelo. In the course of the middle ages, certain tomb-churches in Rome, with a centralised plan, were turned into places of public worship. But, for the plan of the ordinary church, the basilica, with its longitudinal axis, was general. In the eastern empire, on the other hand, the centralised plan was employed from an early date for large churches; and in this way was evolved the magnificent style of architecture which culminated in Santa Sophia at Constantinople. Here the centralised plan was triumphantly adapted to the internal arrangements of the basilica.
Sec. 7. The city of Ravenna, closely connected historically both with Rome and Constantinople, contains a series of monuments which is of unequalled interest in the history of the centralised plan. (1) The mausoleum of the empress Galla Placidia, sister of the emperor Honorius, who died in 450 A.D., is a building of cruciform shape, consisting of a square central space covered by a dome, with rectangular projections on all four sides. The projection through which the building is entered is longer than the others, and the plan thus forms the Latin cross so common in the churches of the middle ages. (2) To the same period belongs the octagonal baptistery, known as San Giovanni in Fonte. (3) In 493 A.D. Theodoric the Ostrogoth obtained possession of Ravenna. To the period of his rule belongs the Arian baptistery, also octagonal, known as Santa Maria in Cosmedin. (4) Theodoric died in 526 A.D. His mausoleum is formed by a polygon of ten equal sides, with a smaller decagonal upper stage, a circular attic above which bears the great monolithic dome. In the lower story was the tomb: the internal plan is a Greek cross, i.e. there is a central space with recesses of equal depth on all four sides. (5) In the year of the death of Theodoric, the octagonal church of San Vitale was begun. It was consecrated in 547, when Ravenna had become the capital of the Italian province of Justinian's empire. Its somewhat complicated plan was clearly derived from an eastern source, but not from Santa Sophia, which was not begun till 532 A.D. The central space is almost circular. Between each of the piers which support the octagonal clerestory at the base of the cupola is an apsidal recess, with three arches on the ground floor opening into the encircling aisle, and three upper arches opening into the gallery above the aisle. On the east side of the central space this arrangement is broken, and one tall arch opens into the chancel, which ends in a projecting apse, semi-circular inside, but a half octagon outside. The aisle with the gallery above thus occupies seven sides of the outer octagon, the eighth side being occupied by the western part of the chancel.
Sec. 8. Of the two types of plan, which can be studied so satisfactorily at Ravenna, the ordinary basilican type is the more convenient. The long nave provides the necessary accommodation for worshippers, the raised apse gives a theatre for the performance of service within view of everybody, the aisles facilitate the going and coming of the congregation, and prevent over-crowding. The centralised plan provides, it is true, a large central area conveniently near the altar; but the provision of a chancel or altar-space necessitates the grafting on the plan of a feature borrowed from the ordinary basilica, which, as at San Vitale, breaks the symmetry of the design. At Santa Sophia, the basilican chancel forms an indissoluble part of a centralised plan; but this feat is beyond the reach of an ordinary architect. Even at San Vitale the planning is highly complicated, and must be due to an architect of some genius. In addition to complications of design, the centralised plan raised questions of roofing which did not trouble the builders of the long wooden-roofed basilicas. The vaulted half-dome of the basilican apse was a simple matter, compared with the mighty dome of Santa Sophia and its cluster of abutting half-domes. It was in the centralised churches, with their domed vaults and the groined vaults of their aisles, that the history of medieval vaulting began. But, even when medieval masons had learned to regard the vaulting of their churches as the controlling principle of their art, they left the centralised plan almost entirely alone, and applied what it had taught them to the work of roofing basilicas with vaults of stone. We shall trace the influence of the centralised church as we proceed; but the influence of the basilica will be found to predominate in the history of medieval planning.
Sec. 9. In England, as in other portions of the Roman empire, we might naturally expect to find the basilican plan applied to the earliest Christian churches. The foundations of a small Romano-British basilican church have been discovered at Silchester in Hampshire. The apse, as in the Roman basilicas, was at the west end. The nave had aisles, which, at the end nearest the apse, broadened out into two transept-like projections. The entrance front of the church was covered by a narthex, the whole width of nave and aisles. This feature, as has been shown, is of eastern rather than of Roman origin; while the projections at the end of the aisles appear to have been, not transepts like those at old St Peter's, but separate chambers corresponding to those which, in eastern churches, flank the chancel, and are used for special ritual purposes. In fact, the basilica at Silchester recalls the plans of the early basilicas of north Africa more closely than those of the basilicas of Rome; while it has, unlike them, the Roman feature of the western apse. This, however, gives rise to questions which, in our present state of knowledge, are beyond solution.
Sec. 10. Of the seven churches which are usually connected with the missionary activity of St Augustine and his companions, five, of which we have ruins or foundations, certainly ended in apses; and the apse in each case was divided from the nave, not by a single arch, but by an arcade with three openings, which recalls the screen-colonnade at old St Peter's. But only one church in the group, the ruined church of Reculver, followed the plan of the aisled nave of the basilica. From the description which remains of the early cathedral of Canterbury, destroyed by fire in 1067, we can see that it, too, was an aisled basilica, with its original apse at the west end. But the first cathedral of Rochester, the plan and extent of which may be gathered from existing foundations, was an aisleless building with an eastern apse. The church of St Pancras at Canterbury, the lower courses of the walls of which in great part remain, had an aisleless nave, divided from an apsidal chancel by a screen-wall with three openings, that in the middle being wider than the others. The foundations of two of the four columns which flanked these openings can still be traced. The walls of the chancel, which was slightly narrower than the nave, were continued straight for a little way beyond the screen-wall; and then the curve of the apse began. St Pancras also possessed a square entrance porch, much narrower than the nave, at its west end, and two chapels projecting from the nave on either side, half-way up its length. The church is thus cruciform in plan. The western porch and the chapels seem to have been added as the work proceeded, and not to have been contemplated in the original design. The material of the building is Roman brick, and buttress projections occur at the western angles of the nave and porch, in the fragment which remains of the south wall of the chancel, and at the outer angles of the side chapels. Small buttresses are also found at the angles and on the sides of St Peter's on the Wall in Essex.
Sec. 11. In one respect the plan of St Pancras at Canterbury is allied to that of the church at Bradford-on-Avon in Wiltshire. At Bradford there remains one of the two porches, which also were probably side chapels, projecting from the sides of the nave. But at Bradford the remaining porch is larger in proportion to the nave than is the case at St Pancras. There is no entrance porch on the west side. Further, the chancel at Bradford is rectangular, not apsidal. Instead of a screen-wall with a central opening nine feet wide, the wall dividing nave from chancel is pierced by a small arch only 3 ft. 6 in. wide. The date of this little church is a matter of great difficulty; and the character of its masonry seems to demand for it a later date than the early one popularly claimed for it. The contrast with St Pancras is accentuated further by the fact that the internal measurements of the nave show a different scheme of proportion. The nave of St Pancras is some three feet broader in proportion to its length than the much shorter nave at Bradford.
Sec. 12. A closer parallel to Bradford-on-Avon is found in the little church of Escomb, near Bishop Auckland. No record of the early history of this building is known; but its masonry is almost entirely composed of re-used Roman dressed stone-work. In this respect it presents a contrast to Bradford. In another respect the two churches are unlike. Both have their entrances in the side walls; but at Escomb there were no original porches covering the doorways, while there are traces of what may have been an entrance porch, like that of St Pancras, at the west end. But they have these points in common: (1) the nave at Escomb is long in proportion to its width; (2) the chancel is a rectangular eastern projection, narrower and much shorter than the nave; (3) there is a solid wall of division between nave and chancel, pierced by a narrow arch, broader than that of Bradford, but very much higher in proportion to its width. It may be added that the walls of both churches are high in proportion to their length and breadth, and that at Escomb the original windows are small openings with rounded and flat lintel-heads, and with internal splays.
Sec. 13. It is, however, with the plan that we are concerned. We now have met with three separate forms in England, viz. (1) the rare basilican plan; (2) the "Kentish" plan of aisleless nave with apsidal chancel; (3) the plan of aisleless nave with rectangular chancel. We also have seen that the screen-wall is common to (1) and (2), while the single chancel arch belongs to (3); and that side chapels and western porches are found incidentally in (2) and (3). Now, the early date of Escomb, apart from the evidence supplied by its masonry, can be suspected only by its analogy to the plan of other churches of which the date is practically certain. Two such churches remain in the same county of Durham. One is at Monkwearmouth, now a part of Sunderland. Its nave and the lowest stage of its western tower represent, and in great part actually are, the nave and western porch of an early Saxon church, which is generally identified with the church built here by Benedict Biscop for the monastery which he founded in 672 A.D. The nave was originally aisleless, long, narrow and lofty: the entrance porch had an upper story finished with a gabled roof, and a vaulted ground-floor with entrances on three sides. There was evidently a chancel arch, and probably the chancel was rectangular. The material of the building was not Roman; but, in the decoration applied to it, Roman work was imitated. Only a few miles further north, Benedict founded, in 680 A.D., the sister monastery of Jarrow. The long and narrow chancel of the present church of St Paul was the body of a church somewhat similar to that of Monkwearmouth. Stone-work which may represent the jambs of a broad chancel arch can be traced in the east wall; but this cannot be stated with positive certainty. The lower part of the tower, now between the present chancel and nave, may represent an original western porch; but, in its present state, it is of much later date than the work east of it, and its site must have been broadened when the tower was first planned. At Jarrow there is no Roman stone-work; but one type of Roman masonry has been imitated by the builders in the walls of the chancel, and small decorative shafts, turned in a lathe after the Roman fashion, such as exist at Monkwearmouth, have been found in the building. The inscribed stone, recording the dedication of the church, is preserved in the wall above the western tower-arch: the date given is 23 April, 684 A.D. In this inscription the building, though aisleless, is called a basilica. The word was now probably used to signify a Christian church, irrespective of its plan. A third early church in this district is that of Corbridge, near Hexham. Here, as at Monkwearmouth, the ground story of the tower was originally a western porch; while the lofty arch between tower and nave is, like the chancel arch at Escomb, entirely composed of dressed Roman masonry, and seems to have been removed from one of the buildings of the Roman station of Corstopitum, as the arch at Escomb was probably removed from the not far distant station of Vinovium.
Sec. 14. The date to which these four northern churches may be assigned is the half century of the activity of St Wilfrid in England (664-709 A.D.). Bede's account of the architectural work of Wilfrid's friend, Benedict Biscop, shows that he procured, for the building of the church at Monkwearmouth, stonemasons and glaziers from Gaul, who were acquainted with "the manner of the Romans." The account which another contemporary, Eddius, gives of Wilfrid's church at Hexham, is clear proof that this important building was a reproduction, in plan and elevation, of the aisled basilicas of the continent—a fact in keeping with Wilfrid's life-long aim of bringing English Christianity into closer touch with the main current of historic Christianity in Rome and Gaul. The foundations of the outer walls of most of Wilfrid's church were uncovered when, lately, the new nave of Hexham priory church was begun; but one of its features has been long known, and is of the highest interest. The crypt for relics below the apse and high altar consists of an oblong chamber, with a western vestibule, approached by a straight stairway from the nave. In addition to the western stair, there are two stairs which communicated with the apse. That on the south side remains perfect, and ends in a passage and vestibule, through which the relic-chamber is entered. The northern stairway leads through a passage to the western vestibule, at the foot of the stair from the nave. The crypt of Wilfrid's contemporary basilica at Ripon also remains: here the arrangement is less complicated; but the arrangement of the main relic-chamber is equally the chief feature of the plan.
Sec. 15. The foundations of the Saxon church at Peterborough present many difficulties, and may be of a later date than the foundation of the monastery in 655 A.D. But no such difficulties of date or plan exist with regard to the large Saxon church at Brixworth, between Northampton and Market Harborough. Its size and the fact that Roman material has been much re-used in its building have given rise to the tradition that it is a secular basilica applied to the purposes of a Christian church. As a matter of fact, the Roman brick-work has been re-used in obvious ignorance of Roman methods; so that this circumstance alone would make the legend improbable. The date of the building can hardly be earlier than about 680 A.D., when a monastery was founded here by a colony of monks from Peterborough. The plan originally consisted of (1) a western entrance porch, with a lofty western doorway, and smaller doorways on north and south; (2) a broad nave, divided from the aisles by arches, which spring from large square piers of plain brick-work; (3) a rectangular presbytery, divided from the nave by a screen-wall pierced with three arches; (4) an apsidal chancel, entered from the presbytery by a single arch. On each side of the chancel arch, a doorway entered into a narrow vaulted passage below the ground level, which probably formed an aisle round a crypt below the apse. At a later date, probably in the period of quiet following the later Danish invasions, the apse seems to have been rebuilt, polygonal externally, semi-circular on the inside, and the central crypt-chamber was then possibly filled up. The western porch was also used as the foundation for a tower, and the western arch blocked up with a filling containing a lower doorway, through which the circular turret for the tower-stair was entered. The aisles, either then or at a somewhat later date, having probably fallen into ruin, were removed. The clerestory of the nave remains, with unusually broad round-headed windows.
Sec. 16. The original plan of Brixworth has points in common with some of the other plans which have been noted. In its triple arched screen-wall it recalls the Kentish type of church; its rectangular presbytery between nave and apse is a development of the chancel space which existed west of the spring of the apse at St Pancras. It shares its western porch with St Pancras and two, if not four, of the northern group of churches. In the north and south doorways of this porch it has kinship with Monkwearmouth, and at Brixworth there are definite signs that these doorways led into passages which may have been connected with other buildings of the monastery, or possibly even with an atrium or fore-court. The aisled nave and the traces of a crypt bring it into relation, not merely with Hexham or Ripon, but with the historical church plan of western Europe generally. At the same time, the plan, regarded as that of an English church, is exceptional. The aisled plan of the parish church was arrived at in spite, not in consequence, of the few early aisled churches which might have supplied it with a model. During the epoch which followed the Danish invasions the aisleless plan was deliberately preferred: the rectangular chancel entirely superseded the apse. No further example of the structural screen-wall occurs. In addition to those mentioned, only three more pre-Conquest examples of crypts are known, and such crypts as occur in parish churches after the Conquest are exceptional, and are usually due to exigencies of site. Only three more aisled churches of unquestionably pre-Conquest date exist above ground. Reculver has been mentioned. The others are Lydd in Kent, where only indications of an arcade remain, and the complete basilican church of Wing, near Leighton Buzzard, which has a polygonal apse with a crypt below. Wing is probably much later in date than most of Brixworth, but one cannot but be struck by a certain resemblance in construction between the two naves, and in plan between the crypt at Wing and the remains of the crypt at Brixworth.
Sec. 17. These early churches have been treated at some length, because they contain certain essential elements of planning in a state of probation. The basilican plan was doubtless the ideal of English builders during the sixth and early seventh centuries, but an ideal which was hard to compass where good building material was not plentiful. Thus Augustine and his companions contented themselves in most instances with a plan which recalled the aisled basilica, without following out its more elaborate details. It is remarkable that they should have departed from the usual Roman custom, and made their chancels at the east end of their churches: it is also remarkable to find at St Pancras the western porch, the origin of which appears to be the non-Roman narthex. Models existed, no doubt in the ruins of the Romano-British churches, which they repaired; and we have seen that at Silchester there is a regular narthex, while, on the other hand, there is a western apse. These models, however, were probably all of one general type, in which the chancel end was formed by an apsidal projection. When Roman Christianity reached the north, it had to contend with the efforts of Celtic missionaries; and those efforts were not met by it effectively until, in 664, the energetic leadership of Wilfrid secured a triumph for his party at the council of Whitby. Of the Celtic churches of the north we know but little: it seems likely that they were for the most part plain oratories of stone or wood, with or without a separate chancel. The simplest form, obviously, which a church can assume is a plain rectangle with an altar at one end. As the desirability of a special enclosure for the altar is recognised, a smaller rectangle will be added at the altar end of the main building, and so the distinction between nave and chancel will be formed. There are indications of this natural growth of plan in some of the early religious buildings in Ireland. In remote districts, as in Wales, the simple nave and chancel plan is general all through the middle ages; and the smaller country churches often follow the common Celtic plan of a single rectangle with no structural division. The ruined chapel at Heysham in Lancashire, a work of early date, is an undivided rectangle in plan. This is the form which would suggest itself naturally to the unskilled builder: the division of nave and chancel into a larger and smaller rectangle is the next step which would occur to his intelligence in the ordinary course of things. It is possible that Wilfrid and Benedict Biscop found that their aims would be best served by adhering in certain instances to the familiar Celtic plan, and so, while they hired foreign masons and craftsmen to build and furnish their earlier churches, and to set the example of building stone churches after the manner of the Romans, they were careful to avoid the prejudice which insistence on a new plan would have excited. The simplicity, moreover, of a plan like that at Escomb, which requires little architectural skill to work upon, may have been a recommendation; and the fact that the construction of an apse is more difficult than that of a rectangular chancel must have weighed powerfully with English masons, both at this time and later. The fact remains that, in the early age of our church architecture in stone, the aisled basilica was a rare exception, and the rectangular chancel was, in the north, at least as common as the apse.
PARISH CHURCHES OF THE LATER SAXON PERIOD
Sec. 18. In later Saxon churches the aisleless plan and the rectangular chancel were normal. Instances of an aisled plan after the seventh century have been noted already: it has been seen that there are only two definite examples, and, although there may be indications of others, these are few and far between and uncertain. The apsidal chancel again is exceedingly rare. We have noted it in combination with other basilican features at Wing: the instances in which it occurs again are very few, and in these, as in the important monastic church of Deerhurst, there are other variations from the aisleless plan. In by far the largest number of examples, the plan adhered to was that simple one of which we have a complete prototype at Escomb. Late Saxon fabrics which remain free of later additions are few; but there is a considerable number of churches which still keep the quoins of an aisleless Saxon nave in situ, although aisles have been added during the twelfth or thirteenth centuries. Such are St Mary-le-Wigford and St Peter-at-Gowts at Lincoln, Bracebridge in the western suburb of Lincoln, St Benet's at Cambridge, and Wittering, near Stamford. At Winterton in Lincolnshire large pieces of the western part of both walls of the nave were kept as an abutment to the tower, when aisles were added. Sometimes, as at Geddington and Brigstock in Northamptonshire, the whole wall above the nave arcades is the upper part of the wall of the aisleless building; and instances in which blocked window openings, of a not improbably pre-Conquest date, remain in walls that have subsequently been pierced with arcades, are exceedingly common. If an untouched Saxon nave is a rare thing, an unaltered Saxon chancel is obviously rarer. The small rectangular chancel of the large medieval church at Repton, in Derbyshire, is practically unique; it was probably preserved for the sake of the crypt beneath, which, at first a plain rectangular chamber, was subsequently, but still in pre-Conquest times, vaulted in compartments supported by columns. But at Sidbury in Devon, where there is a small rectangular crypt, the chancel above was rebuilt in the twelfth, and lengthened in the thirteenth century, without any reference to the line of the walls of the crypt below it. A good example of an unaltered late Saxon fabric is the church of Coln Rogers in Gloucestershire. Here the western tower, built up inside the nave, is a later addition, but the nave, rectangular chancel, and arch between them, are still intact. The chancel arch, though by no means broad, is yet much wider than those at Escomb and Bradford-on-Avon; and its width probably represents the normal width of a chancel arch of this period.
Sec. 19. An addition occurs in most of these late Saxon plans, which had a great influence on the subsequent, and even on the contemporary, development of the church plan. We have noted that at Rome and Ravenna towers formed no part of the original basilican plan, but were added later as campanili. In England it appears that the tower formed no part of the plan until, at any rate, the epoch of the Danish wars.
Western bell-towers were very general by the beginning of the eleventh century. In most of these towers, the ground floor forms an entrance porch; but it does not follow that the western tower in England was generated by the heightening of the western porch. The porches of Brixworth and Monkwearmouth were probably not heightened until the western tower had come into existence elsewhere. An origin for the western tower has been sought in the fore-buildings which occur in some of the early German churches, and contain separate upper chambers. It may be that, derived from this source, the western tower superseded the porch, and, where porches existed, they were adapted to the new fashion.
Sec. 20. The towers of Earl's Barton, Barnack, and St Peter's at Barton-on-Humber, are perhaps the most obviously interesting relics of Saxon architecture which we possess. All are much larger in area than the normal western tower of the later Saxon period. Earl's Barton is a western tower, and its ground floor has probably always served as a porch: the rest of the church, however, is a medieval building of various periods. At Barnack, again, the complete plan of the Saxon church has been lost. Here, however, the western tower was something more than a porch. The doorway is not in the west, but in the south wall; and in the west wall, inside the church, is a niche with a triangular head, which was certainly neither doorway nor window, but a seat. Whether this implies that the ground floor of the tower was used for special religious functions, or for some purpose connected with the common life of the parish, is not clear; but it shows, at any rate, that there was some good reason for the unusually roomy planning of the tower. We stand on firmer ground at Barton-on-Humber. Here, again, a large medieval church exists to the east of the tower. But upon its western side is a small rectangular building of contemporary date, which was not a porch in front of the tower, but a westward extension of the body of the church, the main entrances being on either side of the tower. The foundations of a similar projecting building have been discovered to the east of the tower, beneath the floor of the later nave. It is therefore clear that the ground floor of the tower, or rather of a high tower-like building, formed the body of the church, and that the eastern projection was the chancel. There are clear indications at Broughton, also in north Lincolnshire, that this plan was used, at any rate, once again. The tower at Broughton is obviously later than that at Barton: the doorway, whose details are of a post-Conquest character, is in the south wall; and a large circular stair-turret, like that at Brixworth, projects from the west wall. Probably there was only a chancel here, and no western annexe to correspond. A similar stair-turret occurs at Hough-on-the-Hill, between Grantham and Lincoln: the tower, now western, has a doorway in the south wall, and probably stands mid-way in date between Barton and Broughton. It is planned on a very ample scale, with thin walls and a large floor-space. The main fabric of the church is altogether of a later date; and there are no indications, at any rate above ground, of an earlier building east of the tower. The size of the tower, the provision of a stair-turret, as at Broughton, to leave the ground floor clear, suggest that here we may have a third example of the plan in which the tower covered the main body of the church. The arrangement at Barnack gives grounds for a suspicion of something of the same kind there. In all these cases the tower has been a tower from the beginning; but at Barton-on-Humber the uppermost stage was added towards the end of the Saxon period.
Sec. 21. In these buildings we seem to discover the influence of the centralised plan, acting through the channel of German art. It would be absurd to say that the plan of Barton-on-Humber was inspired by the plan of the palace-church at Aachen, which was an adaptation, with some improvement, of the plan of San Vitale at Ravenna. No masterly intellectual effort, such as the Aachen plan shows, was necessary to plan a rectangle with two smaller rectangles at either end. But the church at Aachen had made the centralised plan familiar to the builders of western Europe. In Germany and in France there are traces of its influence; and we may reasonably suppose that the builders of Barton-on-Humber were acquainted with the existence of an alternative to the usual plan of the church with a longitudinal axis, and did not arrive by haphazard at their concentration of the plan upon a central point. One earlier example of the centralised plan is known to have existed in England. In addition to his basilica at Hexham, Wilfrid had built another church there in the shape of a Greek cross. The description of it which we possess shows that the central space was the actual church, that it was tower-like in form, and nearly circular in shape, and that the arms were simply porch-like projections. Probably it was a combination of baptistery with tomb-church. It is not likely that the simple plan of Barton was derived from that at Hexham. Both were probably the result of continental influence; but, while the church at Hexham may have been the work of Gallo-Roman masons in direct communication with the general current of architectural progress, the church at Barton was probably built by Englishmen, who adapted the centralised plan to methods natural to their comparative want of skill.
Sec. 22. Neither at this time nor later did the centralised plan in England develop along the lines suggested by Barton-on-Humber. No real development on such lines was possible. In Germany, the achievement at Aachen made possible the polygonal nave of St Gereon at Cologne and the centralised plan of the Liebfrauenkirche at Trier, as well as many twelfth and thirteenth century churches whose complicated parts are planned and massed together with relation to a central tower space. In England, however, the habit of dealing with circular or polygonal forms made little progress; and our few "round churches," the plan of the naves of which was a devout imitation of the church of the Holy Sepulchre at Jerusalem, and our polygonal chapter houses, are almost all that we have to show in the way of attempts at a definitely centralised plan. Our church plan develops as the result of an effort to combine a series of rectangles effectively; and, while this combination can be attempted in several different ways, it is obvious that the rigid lines of the rectangle do not admit of that free scope in centralised planning which is given by the circle or polygon.
Sec. 23. We have seen, however, that, even in the earliest days, there was a tendency to admit additions to the simple longitudinal plan, which, in process of time, were bound to give birth, if not to a definitely centralised plan, to something, at any rate, in which a central point counted for much. A feature of the early cathedral and of St Pancras at Canterbury, was the projection of porticus, porches or side chapels, from the nave. These were entered by archways pierced in the centre of the lateral walls. In the cathedral they had outer doorways, and formed the main entrances of the church, on the north from the monastery, on the south from the city. The south porch contained the altar of St Gregory, and, as Eadmer tells us, was used as a court of justice to which litigants, in process of time, resorted from every part of England. In the north porch, dedicated to St Martin, was held the school of the monastery. Upon both porches towers were built at a date which cannot be ascertained, but was probably later than the time of Augustine. Of the use of the porches at St Pancras, which did not contain outer doorways, it is impossible to say anything definitely. Entrance porches, of which one remains, projected from the sides of the church at Bradford-on-Avon: the outer and inner doorways of the north porch are extremely narrow, and are placed west of the centre of its north and south walls. It is possible, therefore, that there was an altar in this porch, so that it served the double purpose of entrance porch and side chapel.
Sec. 24. As time went on, the western porch beneath the tower was disused as a public entrance. The principal entrance of most churches is on the south side, west of the centre of the aisle wall, and is usually covered by a porch. There is a Saxon example of this at Bishopstone in Sussex, where, as at Bradford, room seems to have been left for an altar on the east side. However, the main entrance of the ordinary Saxon church was at the west end, through the ground floor of the tower. The porch in the lateral wall seems to have been regarded primarily as a side chapel; and in some later Saxon churches the porches were dissociated from lateral doorways, and were planned as closed projections from the eastern part of the north and south walls of the nave. This seems to have happened at Britford, near Salisbury, where archways remain on both sides near the east end of the nave. At Deerhurst square projections were entered from both sides of the nave, immediately west of the chancel arch; and it is probable that there were somewhat similar projections at Repton. At Worth in Sussex, where the north and south doorways of the nave are Saxon, and there is no western entrance or original tower, there are large Saxon chapels projecting from the eastern part of the nave, and entered by wide arches. The cruciform plan is sufficiently marked in the conjectural restorations of Deerhurst and Repton. At Worth it is quite unmistakable.
Sec. 25. At Worth, however, in spite of the dignity of the lateral arches, the chapels are still porch-like excrescences, larger in scale than usual, but lower in elevation than the nave. In elevation their transept-like appearance is less noticeable than on plan. Moreover, the length of the nave remains unbroken from west wall to chancel arch: no central space is marked off to which these transeptal projections give emphasis. Nevertheless, a suggestion of an intermediate space between nave and chancel is given; and this space is definitely marked in the plans of churches which may be quite as early in date as Worth—i.e. about the first half of the eleventh century—by the admission of a tower between nave and chancel. The eastern part of the walls of the nave at St Mary's in Dover Castle are continued upwards as a tower, with small rectangular chapels projecting from the sides of the ground floor. Externally, no division between the tower and nave is noticeable; but, inside the church, in addition to the chancel arch and the arches into the chapels, a fourth arch is pierced in the western wall of the tower, and so an intermediate space between tower and nave is effectually created. At Breamore in Hants, a further step is taken. The tower space, between nave and chancel, is of the same width as the nave; but, in addition to the necessary internal division, an external division is also marked by the quoins of the tower, which are complete to the ground. Only one chapel remains at Breamore, on the south of the tower, entered by a narrow Saxon archway; but there was originally another on the north.
Sec. 26. The chapels which project from these early "central" towers are, it is to be noted, not true transepts. They are narrower than the tower, which is built up from the ground, and not upon a system of piers and arches which require lateral abutments in the form of transepts. The western tower is transferred, as it were, to a point near the centre of the church, assumes the width of the nave, and is provided with transeptal excrescences, to communicate with which its side walls are pierced. Such excrescences are not necessary. At Stanton Lacy, in Shropshire, there is only one. At Dunham Magna, in Norfolk, and other places, such as Waith in Lincolnshire, there are, or were originally, none at all. The construction of the "central" tower upon piers connected by arches was beyond the skill of the ordinary Saxon builder; and its natural consequence, the development of the full cruciform plan, with transepts of the height and width of nave and chancel, was thus out of his reach. We know, from contemporary evidence, that one important abbey church, that of Ramsey, had a central tower which was built upon piers and arches as early as 974 A.D.; and perhaps this was the case in other large churches. But, even in the large church of Stow in Lincolnshire, which is commonly taken on trust, without sufficient historical evidence, as the cathedral church of the Saxon diocese of Lindsey, although an advance in transeptal construction was made, the main principle was imperfectly grasped. This church was made the home of a community of clergy about the beginning of the reign of Edward the Confessor, by Leofric, earl of Mercia, and his wife Godiva. It was restored after the Conquest by Remi, the first Norman bishop of Lincoln. The aisleless nave and chancel are Norman work of two periods: probably the nave was rebuilt upon Saxon foundations. The transepts, however, of considerable length and equal height with nave and chancel, were retained from the pre-Conquest building. The tall jambs of the arches of the central tower also remain on all four sides. The arches which they bear are of early Norman character; and the present tower is a late Gothic structure, the arches and piers of which are built up on the inner side of the older masonry. But the Saxon tower space, including the area of the arch-jambs, is rather wider than the arms of the cross which project from it. The tower formed a separate building, with quoins complete from the ground, and nave, chancel, and transepts, instead of combining to support it, were mere excrescences from it, entered by arches in its walls. Possibly the example of Barton-on-Humber may have had to do with this treatment of the tower as a separate central pavilion, which may have been deliberately preferred to the arch and pier treatment. In other respects the plan is an advance upon the plans of Dover and Breamore. And the necessary advance upon Stow is found in the church of Norton-on-Tees in south Durham. Here the tower, between nave and chancel, rests on piers connected by arches. The arches have been widened; two have been entirely rebuilt at a later date; and the rest of the church has been subjected at different times to enlargement and rebuilding. In spite of this, we have at Norton our earliest surviving example of a plan in which the various portions of the church—nave, chancel, and transepts—are gathered together in one structural connexion. The tower is to the east of the centre of the longitudinal axis of the church; but structurally, it is the central point with regard to which the building is planned, and the unity of the composition depends upon it.
Sec. 27. We have arrived thus at a centralised plan of cruciform shape, of which the component parts are rectangular, the central space being approximately a square. The examples which have been given cannot be proved to follow one another in chronological order, but they represent successive steps in planning and construction, of which Norton-on-Tees is the highest. The importance of the inclusion of the tower in the plan is obvious. In its early appearances, its position is unsettled, but the natural tendency is to place it above a main entrance; and this is usually at the west end of the building. Where the builders aim at a simple centralised plan, the high central rectangle will form, like the round or octagonal central space of Wilfrid's church of St Mary at Hexham, ecclesia ... in modum turris erecta, and, as at Barton-on-Humber, will possibly be heightened by a later generation into a real tower. The distinction of the side chapel from the entrance porches, becoming more fully recognised, will lead to the building of transeptal chapels at the east end of the nave; and thus an important addition will be made to the ordinary longitudinal plan. The need of some central building, against which these additions may abut, will be felt. The tower will thus be introduced between nave and chancel, either as an independent structure, or as an upward extension of part of the side walls. The transepts thus, as at Stow, can be raised to an equal height with nave and chancel. From this to a plan in which the component parts are recognised as interdependent, and are closely knit together in structural unity, is an obvious step. At this point, architectural skill, as distinct from mere building ingenuity, comes into play.
Sec. 28. As we proceed, we shall find survivals of old plans, even at an advanced period in the middle ages, which prove that progress in architecture was by no means of an uniform kind. Builders in remote, and especially in hilly, districts, from Saxon times to the present day, have naturally restricted themselves to plans which require as little cost as possible to carry out. Local building material is also an important consideration. In districts where good building stone is to be obtained on the spot, or where money is plentiful and water carriage is possible, the development of plan is naturally rapid, and every fifty years or so, additions to churches will be made in which the old plan will become entirely transformed. In woodland districts, the plan will be controlled to no small extent by the requirements of timber construction. In such regions, Saxon churches were probably built of wood. The only wooden church of Saxon times which remains is that of Greenstead in south Essex, with a rectangular chancel and aisleless nave constructed of vertical logs placed side by side, and framed originally into a timber plinth. However, it may be stated as a general rule, that, whatever may be the helps or hindrances to development provided by local materials, the real starting-point of the parish church plan of the middle ages is in every part of the country an aisleless plan; and that this plan consists either of a nave and chancel with a longitudinal axis, or of a nave and chancel whose longitudinal axis is intersected by a transverse axis across transepts. Variations, no doubt, occur; but these will never carry us far from one or other of these fundamental plans. The aisled basilica of the continent found no scope for itself in Saxon England; and it was through an interval of aisleless building that the aisled plan eventually became acclimatised, and then in a form which bears only a superficial kinship to the basilican plan.
THE AISLELESS CHURCH OF THE NORMAN PERIOD
Sec. 29. During the century after the Norman Conquest, the great abbey churches and cathedrals represent the work of a foreign architectural school, gradually acclimatising itself in England; while, on the other hand, the parish church continued to be planned by local men, open to receive the improvements which more skilled foreign masons had introduced. Consequently, while local art received a continually increasing refinement, the plan of the church developed upon traditional lines, and not upon those novel lines which foreign masons would have laid down for it. The chief proof of this is seen in the persistence of the aisleless plan with rectangular chancel and western tower. The tendency of a Norman builder would be to design his church with an apsidal chancel, transepts, and a central tower; his practice would vary, but this would be his favourite plan. On the other hand, the rectangular chancel and western tower remained the favourite terminations of the parish church in England. But, while a large number of rubble-built, unbuttressed Norman towers, usually heightened or otherwise altered in the later middle ages, remain in many parts of England, their relation to the plan suffers some change. The ground floor of the Saxon tower was, as we have noticed, the main entrance to the church. The Norman western tower either contained no western doorway at all, or provided merely an entrance, which was used only on special occasions. At Caistor the ground floor was probably the main porch of the aisleless church; and there are exceptional instances, as at Finchingfield in Essex, where, in fairly advanced Norman work, the same arrangement was clearly contemplated. On the other hand, at Laceby, between Caistor and Grimsby, a south doorway, coeval with the western tower, has always been the main entrance to the church. Similarly, at Hooton Pagnell, and at Blatherwycke in Northamptonshire, south doorways, of the same age as the tower, form the chief entrance. These last three are early Norman examples; but we may go back even further, to find the same thing in churches which are usually reckoned as late Saxon work, at Heapham in Lincolnshire, and Kirk Hammerton, between York and Boroughbridge. In south Yorkshire there are a few churches of the middle of the twelfth century whose western towers are noticeably derived, in their plan and general construction, from the Saxon type—Birkin, Brayton, and Riccall. But in all three, the main entrance to the church was made through a south doorway, the arch of which is covered with elaborate late Norman ornaments. The western tower was thus reduced to the state of a bell-tower at one end of the church, and, while increasing in size and in magnificence, was actually a less indispensable part of the plan than before.
Sec. 30. The nave of the Norman aisleless church was usually short, and, where the church was entirely rebuilt, rather wide in proportion to its length. The naves of churches like Garton-on-the-Wolds or Kirkburn in Yorkshire, give the effect of spacious halls, of no great length, but wide and lofty. It cannot be doubted, however, that the fabric of the Saxon church was frequently kept, or that the church was rebuilt upon Saxon foundations. It is not unusual, as already stated, to find Saxon quoins still existing at the angles of naves to which aisles have subsequently been added. Evidences, on the other hand, of the westward lengthening of a Saxon nave in the Norman period appear to be rare. At North Witham in south Lincolnshire, the south and (blocked) north doorways are Norman work, in the usual position near the west end of the nave. East of them, however, in the centre of the nave walls, there are distinct traces of the inner openings of a north and south doorway, which may belong to the late Saxon period. That we have here a case of the twelfth century lengthening of an earlier nave may be inferred. The probability is increased by the fact that, in the neighbouring church of Colsterworth, where aisles were added during the early Norman period to a late Saxon fabric, the nave and aisles, towards the end of the twelfth century, were certainly extended a bay westward. As little architectural work is done without a precedent, we may assume that the builders at Colsterworth were following the example of North Witham.
Sec. 31. The great majority of Norman rectangular chancels have been lengthened and enlarged; for the plain "altar-house" at the east end of the nave was too small for the purposes of the ritual of the thirteenth and fourteenth centuries, and afforded no intermediate space between nave and chancel. However, short and approximately square chancels were by no means invariable; and, before the middle of the twelfth century, oblong chancels of considerable length in proportion to their width were being built. There is a good early twelfth century example at Moor Monkton, in the Ainsty of York; and the chancel of the middle of the twelfth century at Earl's Barton, Northants, is of considerable depth, and was of ample size for all later purposes. At Earl's Barton the eastern portion was the chancel proper; while the western portion supplied that space for a quire which was not provided in less elongated plans. In by far the larger number of cases, the rectangular chancel had a wooden roof. There is, however, a fair number of churches in which the system of ribbed vaulting, as employed in larger buildings, was used. Thus at Heddon-on-the-Wall, Northumberland, there is a small square chancel with a ribbed vault. At Warkworth, there is a long vaulted chancel of two bays, built during the first quarter of the twelfth century; and at Tickencote, Rutland, two bays are combined in one by the use of sexpartite vaulting. In these cases the chancel arches are wide, forming the western transverse arches of the vaulting: that at Tickencote is of remarkable magnificence.
Sec. 32. There are certain cases in which the chancel was of the same width as the nave, and no structural division existed between them. At Askham Bryan and at the chapel of Copmanthorpe, near York, the plan, externally and internally, is a plain undivided oblong. At Tansor, Northants, the chancel was rebuilt about 1140, when the side walls were set back in a line with those of the nave. In St Mary's in the Castle at Leicester, the long and very narrow nave was, as may still be clearly seen, continued eastward without a break into the long and narrow quire and chancel. Here the eastern half was used, no doubt, by the college of dean and canons, while the western half was the parish church. The beautiful church of St Peter, Northampton, built towards the end of the third quarter of the twelfth century, gives us a complete example of an undivided plan, aisled throughout save in the eastern bay, which forms a projecting chancel east of the aisles of the choir.
Sec. 33. Hitherto we have dealt merely with the rectangular chancel. But there are also churches which end in an eastern apse. These are comparatively few and exceptional. In Yorkshire, where the number of Norman rectangular chancels is large, and buildings such as Adel exhibit the aisleless church in its highest state of architectural development, the number of apsidal chancels can be counted on the fingers of one hand. In Sussex, where Caen stone was largely used, and we should expect foreign influence to be noticeable, the proportion of apsidal chancels is small. In Gloucestershire, the Cotswold district contains several small Norman churches, which have been little altered: the rectangular chancel is universal. These are typical districts; and, to state a general rule, we may say that, while the apsidal chancel is foreign to no part of England, and occurs in unexpected places, as in the chapel of Old Bewick, Northumberland, it is never general in any single region. Its rarity is an important fact. Were our parish churches the work of masons sent out from the larger churches and monasteries, we should expect to find it a common feature; for in those buildings the apsidal plan prevailed. But, in the hands of local masons, its sparing employment is easily explained. To build an apse needs skill, not only in planning, but in stone-cutting. The question of vaulting the apse increases the difficulty and the expense. These difficulties would not trouble masons who had worked at the building of Durham or Ely or Winchester; nor would expense trouble the monasteries, which, according to the popular idea, were so ready to lavish money on the fabrics of parish churches. Many apsidal chancels have disappeared, no doubt; but, if we take the bulk of those which remain into account, we shall find that they have a habit of occurring in small groups, as in Berkshire, where three occur together within a single old rural deanery, and that the large majority of the churches in which they are found were not monastic property. A few belonged to preceptories of Knights Templars in their neighbourhood; and perhaps we may see in their apses a reference to the circular form of the Holy Sepulchre. But, as a rule, we may say that a band of masons in certain neighbourhoods developed some skill in building apses, that money was forthcoming, and that so a few examples came into existence. In one curious instance, Langford in Essex, which is within easy distance of four or five other apsed churches, there is an apse at the west, and there are foundations of another at the east end of the building. For this church a Saxon origin has been claimed: the plan, at any rate, indicates a survival of a plan once common in western Christendom, and especially in the German provinces. In apsed churches, like Birkin in Yorkshire, the apse does not spring from points directly east of the chancel arch. The arch is wide and lofty; behind it is a nearly square rectangular space, which is divided from the apse by another arch. At Birkin the apse has ribbed vaulting, which allows the walls to be pierced freely for windows. At Copford in Essex, Old Bewick, and other places, the roof is a half-dome without ribs: this allows for the display of mural painting, but admits of less light.
Sec. 34. The most important feature in the apsidal plan is the provision of the distinctly marked quire space between the nave and chancel. This space also occurs in plans where the chancel is rectangular; but in such cases it becomes the ground story of a tower. There are famous examples of this at Iffley, near Oxford, and Studland in Dorset, where the chancels are vaulted. Coln St Denis in Gloucestershire, where the tower is of very wide area, and projects noticeably north and south of nave and chancel; and Christon in Somerset, are further instances of the plan. The tower between nave and chancel, without transepts, is seldom found in an apsidal plan. It occurs at Newhaven in Sussex, where there is a small apse. Here the plan is virtually that of some small parish churches in Normandy, such as Yainville, near Jumieges. The majority of such plans in England, however, end in a rectangular chancel. Precedent for the plan is, as we have seen, to be found in Saxon churches. At St Pancras, Canterbury, we have noticed the westward prolongation of the apse: at Brixworth a definite presbytery or quire space was planned, on a large scale, between apse and nave. In later Saxon churches, where the chancel was rectangular, a tower, with or without transeptal chapels, was sometimes built between nave and chancel; and here, although externally the division was not always clearly marked, an internal quire space was divided off from the nave by the western arch of the tower. The aisleless plan, therefore, with a tower above the quire, and a rectangular chancel, points to a development along old-fashioned lines, even in churches in which, as at Iffley, the builders have acquired great skill in expressing themselves in Norman terms. In certain districts, as in Gloucestershire, this plan was a favourite one. Even in the fourteenth century, Leckhampton church, near Cheltenham, was rebuilt in faithful adherence to this tradition. Here the tower is narrower than the small chancel, and the nave has a south aisle.
Sec. 35. In the cases of Dover, Breamore, Stow, and Norton, we have watched the gradual evolution of the cruciform plan with central tower. It must be noted once more that to the cruciform plan the central tower built on piers and arches is essential. It is possible, as in the Gloucestershire churches of Almondsbury and Avening, to pierce the north and south walls of a tower and add transeptal chapels: the plan will have a cruciform appearance, but will still be only an elongated plan with lateral additions. It is possible, in a church where there is no central tower at all, to extend the side walls at right angles north and south, and so form transepts; but here again the transepts have no structural reference to a central point in the plan, but are mere widenings of the nave or aisles. The thirteenth century aisleless churches of Potterne, in Wiltshire, and Acton Burnell, in Shropshire, are both cruciform in plan. The church at Potterne was planned throughout with reference to the crossing of transepts, nave, and quire, above which its central tower rose: the tower space is the central point of the whole. But, at Acton Burnell, there is no central tower or space: the body of the church consists of a long aisleless nave and an aisleless chancel beyond; and the transeptal chapels are simply stuck on, as it were, to the eastern part of either wall of the nave. This is at once noticeable in elevation, when the chapels are seen to be mere excrescences, with roofs lower than the nave. Moreover, where there is a true central crossing, with a tower above, such as we find in almost all our cathedrals, a transept on either side is necessary for the support of the tower. The transepts need not be wholly symmetrical, although in most cases they are; but they must be there. On the other hand, where there is no central tower, and the crossing is merely apparent, symmetry of treatment is quite unnecessary. While there are two transeptal chapels of similar size at Acton Burnell, or at Achurch in Northamptonshire, there are far more instances in which a less regular treatment was adopted. Thus, at Childs Wickham in Gloucestershire, and Montacute in Somerset, there is only one transeptal chapel, in each case on the north side. At Corbridge in Northumberland, transeptal chapels, extended outwards from the aisle walls, are of different lengths. At Medbourne in Leicestershire, a long aisleless transeptal chapel was built out from the north side of the nave in the thirteenth century. Within the next fifty years a south chapel was built, but, instead of copying the proportions of the northern chapel symmetrically, the builders gave their new chapel a much greater width, and placed its altars in an eastern aisle. The plan is thus accidentally cruciform. At Acton Burnell and Achurch it is, no doubt, designedly cruciform; at Montacute and Childs Wickham, imperfectly cruciform. But all three varieties belong to one class, the longitudinal plan with transeptal extensions. The structural feature which makes the truly cruciform plan, the central tower upon arches and piers, is wanting. And this distinction between churches planned from a centre, and churches whose plan follows a longitudinal axis, although often overlooked, is essential.
Sec. 36. A noble example of a Norman cruciform church, whose plan has suffered little alteration, exists at North Newbald in the east Riding of Yorkshire. At each angle of the crossing are masses of shafted piers, connected by wide and lofty rounded arches. The nave, as is usual, is the longest arm of the four, so that the plan is a Latin cross. It has north and south doorways: there are also doorways in the end walls of the transepts, placed in the western part of each wall. In the east wall of each transept is an arch, now blocked up, the filling being pierced with fifteenth century windows. These arches are the openings of original apses, which contained the transept altars. The chancel, probably always rectangular, was rebuilt in the fifteenth century. As a corollary of the true cruciform plan, the four arms are all of equal width. At Bampton-in-the-Bush, Oxon, where the plan of the church was greatly altered in the thirteenth century by the addition of aisles, the Norman plan was very similar to that of North Newbald. The cruciform plan of Melbourne, Derbyshire, with its aisled nave, was probably inspired more directly by continental examples. The aisleless chancel was vaulted, and ended in an apse, which was squared in later times by the addition of a rectangular piece east of its springing points. Out of the east walls of the short transepts opened wide apses, the walls of which joined the western ends of the walls of the chancel. Thus, externally, the plan of the eastern part of the church was closely allied to the plan with three apses which, in some of our larger churches, was derived from Normandy. At Melbourne, however, there are important variations from this plan. The chancel is short, there are no quire aisles, and the transept apses were rounded externally. In the larger churches of Normandy, the side apses were at the end of the quire aisles, and were usually squared externally, while the apses projecting from the east walls of the transepts, as at Saint-Georges-de-Boscherville, were left rounded. At Newbald and Bampton there seems to have been no attempt to give complete unity of design, as at Melbourne, to the rectangular chancel and transeptal apses. In any case, transeptal apses were the exception in the plans of our Norman cruciform churches, although their convenience for holding altars is obvious.
Sec. 37. The cruciform plan, beautiful as it is, was never generally adopted. It was inconvenient for purposes of public worship, as long as the rounded arch remained fashionable. In our own day, even in churches where the central tower is carried on high pointed arches, and the view of the altar is practically unhindered, the chancel is cut off from the nave by the crossing, and the acoustic problem, which in modern church planning is so necessary a consideration, is almost insurmountable. In the middle ages, this problem was not so acute; but it was undesirable that the interior of the chancel should be nearly invisible from the nave. At Newbald the tower arches are planned upon a liberal scale: at Bampton, on the other hand, where the eastern tower arch is left, the others having been rebuilt in the thirteenth century, it is very low. The low tower arches at Burford, Oxon, and the narrow arches at St Giles, Northampton, are examples of the way in which the supports of the Norman central tower interfered with the internal convenience of churches. It was not until much later that this difficulty was solved, and then only in one or two cases, when the cruciform plan had become exceptional. The plans of Bampton, Burford, and Witney, show how the builders of west Oxfordshire experimented in cruciform planning. The division between chancel and nave is felt much less at Witney than in the other two churches; for the great thirteenth century tower and spire, resting upon massive piers joined by pointed arches, throw a considerable portion of their weight upon nave and transept arcades, whose exceptional massiveness gives unity to the whole design. In the fifteenth century, however, the rebuilders of the aisleless church of Minster Lovell, between Witney and Burford, solved the problem by removing the supports of their square central tower from the angles of the crossing to points entirely within the church, and building arches from the piers thus formed to the angles of the crossing. The comparatively light piers, instead of hindering the view, allow of easy access from the nave to the transepts, and there is hardly a point in the body of the church from which seeing and hearing alike are in any way impeded. With the earlier builders, however, the natural course was to leave the piers where they were, and endeavour to lighten them as far as possible; and, in aisled churches, the difficulties involved often led to the abandonment of the complete cruciform plan.
Sec. 38. The cruciform church gives occasion for a brief remark on one aspect of medieval building which is often exaggerated. The revival of interest in medieval architecture, in the early part of the nineteenth century, was accompanied by an insistence on symbolism in the plan and design of churches. A minute symbolism, which often was the fruit of pious imagination, or was derived from the fancies of post-medieval writers on ritual, was read into every detail of the medieval church fabric. It is true that, as has been said, some builders worked imaginatively, imitating in the round naves of a few churches the rotunda of the Holy Sepulchre. Other instances of devout imitation might be found, if we looked for them. But the imitation of a concrete model is a different thing from translating abstract mysteries into the plan and elevation of a building. And, although the ground plan with nave, transepts, and chancel, certainly forms a cross; and, although, as time went on, the resemblance to the chief symbol of the Christian faith was no doubt recognised and valued, the plan itself, as we have shown, came into being from entirely natural causes. Where the central tower was introduced, the plan was dictated by structural necessity. Where there was no central tower, transeptal chapels provided accommodation for altars, for which the body of the church afforded no convenience. In this and in other cases, medieval builders were impelled by practical common sense and the requirements of the services of the church; and symbolism, if it was a consideration at all, was purely secondary.
THE AISLED PARISH CHURCH
I. NAVE, TOWER, AND PORCHES
Sec. 39. The variations of the aisleless plan, which have been indicated, are all of which it is capable. Naturally, after the twelfth century, many aisleless churches were still built, and are common in country districts. In their humblest form we find them in the small churches of highland regions, the masonry of which is so rough that their date is often a matter of doubt. Sometimes they have been rebuilt, with a lengthened chancel, as at West Heslerton, near Scarborough. In many instances, we have aisleless country churches rebuilt in the fourteenth and fifteenth centuries, with western towers. This, uncommon in no part of England, is especially common in Norfolk and Suffolk; and some of these churches, like Ranworth in Norfolk, have much dignity and spaciousness of proportion. In some late Gothic churches the structural division between nave and chancel is left out, and the building has been deliberately planned as a spacious aisleless rectangle, of which the eastern bay is allotted to the chancel. This happens at Temple Balsall in Warwickshire and the chapel of South Skirlaugh in Yorkshire. Aisleless plans with one or two transeptal chapels are to be found all through the middle ages: Acton Burnell represents a thoroughly symmetrical employment of this type. On the other hand, aisleless cruciform plans with central towers are by no means common after the twelfth century. Potterne is a perfect development of this plan in the thirteenth century. There is a complete aisleless cruciform plan at Othery, near Bridgwater, where the tall central tower is quite out of proportion to the humble church above which it rises, and has necessitated substantial outer buttressing. Here probably the church was rebuilt on earlier foundations, transepts being possibly added. In many instances an aisleless cruciform church seems to have been rebuilt on the lines of a complete Norman plan. This was with little doubt the case at Acaster Malbis, near York, where the church is planned with direct relation to the central space, but without a tower; and the foundations of earlier walls can be traced all round the building, at the foot of the walls built in the fourteenth century. The absence of the tower is an anomaly, but is one method of solving the problem of the connexion between nave and chancel in the cruciform plan.