Skip to main content

Full text of "The arts in early England"

See other formats

' (jr«ruu«r«nHnfuMHnriMrtnftrMif«ritwiiNHMuiiUHniU47^WMM»nt«ifi»iU<Wk,lck^ 




»|il>t<il«Wtlllll W WWII»WM«*W«l*t»yifWWWt>W»M>lf|fMM<IMMfaWHM»WW<Mt>WWWMI4IHW<ltiWW»lt!tW>M"MflN*l«WMi 







— \ 



Western Tower, Earls Barton, Northamptonshire. 











'' ^6 




Prefatory Note 


The present volume is not offered as a history of pre-Conquest 
architecture. Before the history of a style can be written there 
must be some general agreement as to the chronology of its 
monuments. Such a chronology of Saxon buildings has yet 
to be established, and it is hoped that the notices which follow 
of examples of English work, with their continental parallels, 
will be of some aid towards this desirable end. 

But even if the chronology of the existing monuments of 
the period were fixed, a treatment in the true sense historical 
would still be hardly possible. A history implies development, 
and in Saxon architecture, as in Saxon civilization generally, 
there is neither continuous progress nor evolution. In many 
respects the earlier periods of Saxon Christianity, in the North- 
umbria of Ecgfrid or the Mercia of Offa, present a picture 
more attractive than any of the later epochs. Partly owing 
to the Danish desolation, and partly to a tendency in the Saxon 
temperament to sink into that inertia which Bede deprecates 
in his countrymen, there was in Saxon England no continuous 
advance, but rather alternations of brilliant periods and times 
of stagnation or decline. In architecture some of the structures 


of the pre-Danish epoch seem to have been more ambitious 
than anything attempted at later dates. 

It is accordingly inevitable that a treatment of Saxon build- 
ings on the basis of our present knowledge should take the 
form of a descriptive survey rather than a history. The survey 
here offered is a fairly wide one and embraces examples from 
all periods and all parts of the country, while the map and 
index list, giving the names and position of examples, may 
be of service to those who investigate the subject further for 

Previous studies embracing the whole field, but without dis- 
crimination of periods, are contained in Bloxam's Ecclesiastical 
Architecture^ and in the editions of Rickman that ended with 
the sixth. In the seventh, the current edition, the chapter on 
Saxon architecture was withdrawn, and the separate appendix 
that was to take its place has not been issued. Among studies 
of groups of buildings, the papers in The Reliquary of 1893-4 
by Mr. C. C. Hodges on the pre-Conquest churches of the 
ancient Northumbria embrace a larger number of examples 
than any other recent essay, and the writer hereby acknow- 
ledges the assistance he has derived from this source as well 
as from correspondence with Mr. Hodges, who has readily 
opened his stores of knowledge about the early churches of 
the North. There are many published descriptions of special 
buildings in journals of Archaeological Societies, and references 
to some of these will be found in the notes to the text, but 
there has been no effort to form a bibliography of the subject, 
a work which, considering the varied degrees of value in the 
papers that would have to be included, is hardly worth 

Apart from this large body of descriptive material, there is a 


philosophical treatment of examples from the point of view of 
the types they offer in Mr. Micklethwaite's paper in the fifty- 
third volume of the Archaeological Journal. This paper was 
epoch-making in that it introduced for the first time a principle 
of classification among hitherto disjointed units, and it has 
greatly furthered the study of this architectural period. The 
principle of grouping the buildings by types rather than by 
chronological epochs or districts, in our present state of know- 
ledge the only possible one, has been adopted in these pages. 

A considerable part of the matter that follows, with many 
of the illustrations, have appeared in other forms in The 
Builder^ and the writer expresses his thanks to the proprietors 
of that journal for their ready acquiescence in his desire to 
re-issue the matter in the present extended shape, as well as 
for the loan of sundry blocks. Much help has been received 
in certain points from the notes and drawings of pre-Conquest 
buildings bequeathed by the late J. T. Irvine to the Society of 
Antiquaries of Scotland. By the kind permission of the Council 
of the Society and with the assent of Mrs. Irvine a few of these 
drawings have been reproduced in the present volume. 

The plans, with one or two exceptions noticed when they 
occur, are all drawn to the same scale, and as they appear 
on the page are to the scale of one-sixteenth of an inch to 
a foot. The absolute and the comparative dimensions of the 
buildings discussed can be obtained from the plans ; and in this 
connection the subjoined scale of feet may be found useful. 

9 ( 10 It 20 M 30 JS tf> 

I M H 1-4 JH iH H M HI fH M M H H t-l H M (-H ^-^ H M 

Save where otherwise stated the plans are based on the writer's 
own measurements and notes, but for evidence of Saxon work 


not now visible on the sites reference has been made to published 
descriptions and plans, many of which were drawn during periods 
of church restoration when work was uncovered that is now 
again concealed. Space has not permitted a full discussion of 
the details of these plans, in which however only these features 
have been incorporated for which there is good authority. 

The method of treatment adopted in the survey is suffi- 
ciently explained in the text, and it is hoped that the reader 
will find assistance in the cross references that have been 
copiously introduced, as well as in the indices. The citations 
from ' Vol. I.' refer to the historical volume on 'The Life of 
Saxon England in its Relation to the Arts. 

Edinburgh, Maifh, 1903. 




The Roman and the Celtic Sources of English Ecclesi- 
astical Building ------- i 

Foreign Influences on Later Saxon Architecture - 33 


The Number, Distribution and Criteria of the Exist- 
ing Monuments ___-_-_ 70 

The Types and Features of Saxon Churches - - 99 


The Types and Features of Saxon Churches, continued: 

Thj: Western Tower _ - _ - _ ~ 140 




The Types and Features of Saxon Churches, concluded 205 

An Essay in Chronology and History - - - - - 272 

The Architectural Standard of Saxon Buildings - 309 


Index List of Existing Remains and Map of Saxon 

Churches - - - - - - - - -331 

General Index ________ 344 

List of Illustrations 

Western tower, Earls Barton, Northamptonshire - - Frontispiece 

1 . Jamb of Roman gateway, Chesters, on the North Tyne - - 4 

2. Roman doorway at Chesters ----__ j 

3. Roman tooling, Northumberland - - - _ _ _ 6 

4. Roman shaft in Saxon tower, Wickham, Berks - - - 10 

5. Roman dwarf pillar, Housesteads- - - - - - 10 

6. Plan of Early Christian church, Silchester - - - - 11 

7. Plan of church at Guesseria, North Africa - - - - 17 

8. Doorway in old Irish fort - - - - - - - 20 

9. Small square oratory on Skellig Michael, off Kerry, Ireland - 22 

10. Interior of 'Gallerus' --_.___ 24 

11. Oratory on St. Macdara's island - - - - - - 25 

!2. Pilaster at Dulane, near Kells ---___ 25 

13. Sections of stone-vaulted churches . - _ _ _ 26 

14. Plans of churches at Killiney and Glendalough, Ireland - - 28 

15. Interior of Trinity church, Glendalough _ . . _ 29 

1 6. Doorway at Glendalough ------- 30 

17. Details from old Irish buildings - - - - - - 31 

18. Corner of Greenstead church, Essex - - - - - 41 

19. Plan of above --------- 42 

20. Front and side view of the Minster at Aachen - - - 51 

21. Section of west end, Aachen - - - - - - 52 

22. View of gatehouse at Lorsch ------ 59 

23. Base of half-columns from Lorsch ----- 59 



24. Impost from Lorsch ----_-__ 5^ 

25. Mid-wall shaft at Trier _-_-___ 62 

26. Methods of dividing a double opening - - _ - - 64. 

27. Plan of double-splayed light. Trier - _ . _ _ 67 

28. 'Mushroom' capital, Werden a. d. Ruhr _ _ _ . 67 

29. Herring-bone work at Tamworth - - - - - 71 

30. Proportions of Saxon churches ------ 84 

31. Plinths of Saxon walls ------- 85 

32. Quoin at St. Mildred, Canterbury ----- 86 

33. Quoin at Stow, Lincolnshire - - - - - - 87 

34. Long-and-short work, Wittering - - - - - - 88 

35. Earls Barton and West Mailing ------ 89 

36. Pilaster strips at Woolbeding, Sussex ----- go 

37. Pilaster strip, Breamore, Hants ------ go 

38. Tower window at Barton-on-Humber ----- 91 

39. Window in nave at Worth, Sussex - _ _ - _ ^2 

40. Double-splayed window at Diddlebury - - - - 93 

41. Comparative plans of window openings _ - - - g^ 

42. Opening at Brigstock -------- 94 

43. Window at Brigstock with sloping jambs _ _ - _ g_y 

44. Doorway at Diddlebury -- - - - - - 95 

45. Window at Iver, Bucks - - - - - - - 95 

46. Tower arch at Market Overton ------ 96 

47. Chancel arch at Stainton, Yorks ------ 96 

48. Tower arch, Brigstock ----___ gj 

49. Recessed arches with angle shafts ------ 98 

50. South door at Stopham, Sussex --,.-- 98 

51. Plans of single-celled oratories - - - - - - loi 

52. Doorway of Heysham Chapel ------ 102 

53. Doorway at Somerford Keynes - - - - - - 103 

54. Plan of Whitfield, Kent - - - - - - - 104 

55. Plan of Boarhunt, Hants ------- 105 

56. Window at Boarhunt -___--- 105 

57. General view of Boarhunt - - - - - - - 106 

58. Plan of Wittering -._.---_ 107 



59. View and plan of jamb of chancel arch, Wittering - - - 108 

60. Plan of jamb of chancel arch, Clayton, Sussex - - - 109 

6 1 . V^ievv of Deerhurst chapel - - - - - - - 109 

62. Plan of Escomb church, Durham - - - - - - iio 

63. View of Escomb from south-east - - - - - - iii 

64. Interior of Escomb - - - - - - - - 113 

65. North doorway of nave, Escomb, interior view - - - 114 

66. Id., exterior view - - - - - - - - 114 

67. Round-headed window, Escomb - - - - - - 115 

68. Square-headed window, Escomb - - - - - - 115 

69. Plan of Saxon church at Rochester - - - - - 119 

70. Plan of St. Martin, Canterbury - - - - - - 120 

71. Plan of St. Pancras, Canterbury - - - - - - 123 

72. Porch at Bishopstone, Sussex - - - - - - 131 

73. General view of Bishopstone - - - - - - 132 

74. Plan of Bradford-on-Avon - - - - - - - 132 

75. View of Bradford-on-Avon - - - - - - 134 

76. Arcading at Bradford - - - - - - - 135 

77. Arcading at Dunham Magna - - - - - - 137 

78. Plan of Monkwearmouth - - - - - - - 141 

79. Western tower, Monkwearmouth - - - - - 142 

80. Western doorway to porch, Monkwearmouth - - - 143 

81. Sculptured serpents on doorway, Monkwearmouth - - - 144 

82. Turned baluster shaft, Monkwearmouth _ - _ . i^.- 

83. Western wall, Monkwearmouth - - - - - - 147 

84. Plan of Corbridge tower - - - - - - - 151 

85. Western portal, Corbridge - - - - - - - 153 

86. Plan of Bardsey tower, - - - - - - - 156 

87. Western tower, St. Peter-at-Gowts, Lincoln - - - - 160 

88. Tower of ClapJiam church, near Bedford - - - - 160 

89. Arcading at Branston, near Lincoln - - - - - 161 

90. Upper stages of towers - - - - - - - 163 

91. Doorway at Clce, Lincolnshire - - - - - - 164 

92. Keyhole loop --------- 165 

93. Recess in Skipwith tower, Yorks - - - - - - 168 




94. West end of nave, Bosham - - - _ _ 

95. Comparative schemes of western ends - - _ 

96. Plan of Netheravon tower - - - - _ 

97. Plan of belfry, Great Hale - - - - _ 

98. View of Bracebridge by Lincoln - - - _ 

99. Caps of mid-wall shafts - - _ - _ 

100. Moulded imposts ______ 

1 01. Plan of Witton, Norfolk - - - - _ 

102. Plan of Bessingham, Norfolk _ _ _ _ 

103. Plan of Earls Barton tower - - - - - 

104. Western doorway, Earls Barton - - - _ 

105. Plan of belfry stage, Earls Barton _ _ _ 

106. Roman altar from Birrens - - - _ _ 

107. Roman astragal from Hexham - - _ _ 

108. Carved stone from Hexham _ - _ _ 

109. Carved stone from Jarrow - - - - _ 

1 10. Baluster shafts at Jarrow and Monkwearmouth 

111. Fragments of baluster shafts in the Museum at Dover 

112. Baluster from Barton-on-Humber _ _ _ 

1 1 3. Shafts from St. Albans - _ _ _ _ 

114. Baluster from Brixworth - _ _ _ _ 

115. Baluster from Earls Barton - - - - _ 

116. Mid-wall baluster shaft from Wing, Bucks 

1 17. Shaft from Bardsey church, near Leeds - - - 

1 18. Southern jamb of tower arch, Sompting, Sussex 

1 19. Corbel caps from Sompting and Trier - - - 

120. Pierced mid-wall slab, Barnack - - _ _ 

121. Plan of Barnack tower _____ 

122. View into Barnack tower - _ _ _ _ 

123. Barton-on-Humber, general view _ _ _ 

124. Western wall of nave, Barton-on-Humber 

125. Plan of Barton-on-Humber _ _ _ _ 

126. Barton-on-Humber, original form of the Saxon church 

127. Saxon gable cross from Corbridge _ _ _ 

128. Broughton tower, formerly chancel, arch, western face 
















21 1 



129. Broughton tower, formerly chancel, arch, eastern face - - 213 

130. Section of Barton-on-Humber - - - - - - 215 

131. Herring-bone facing at Diddlebury, Shropshire - - - 218 

132. Plan of * Old Minster,' South Elmham, Suffolk - - - 222 

133. Plan at North Elmham, Norfolk - - - - - - 223 

134. View of Dunham Magna, Norfolk _ . _ - - 224 

135. Plan of Dunham Magna - - - - - - - 225 

136. Plan of Saxon portion of Britford church, by Salisbury, Wilts - 228 

137. Face of jamb with springing of southern arch, Britford, Wilts - 229 

138. Soffit of northern arch, Britford, Wilts - _ _ - - 229 
I 39. Eastern part of Repton church, Derbyshire, showing probable 

Saxon plan- - - - - - - - - 23^ 

1 40. Plan of eastern part of Deerhurst, Gloucestershire - - - 233 

141. Plan of Breamore church, Hants - - - - - - 233 

142. Breamore church from the south - - - - - - 234 

143. Archway into south transept or chapel, Breamore, Hants - 235 

144. Plan of Worth church, Sussex ------ 237 

145. Worth church, Sussex, from the south-east - - - - 238 

146. Sketch-plan of central tower and parts adjacent, Stow, Lincoln- 

shire ---------- 240 

147. Saxon cathedral at Exeter, from a seal - - - - - 243 

148. Saxon remains at the north-west corner of Lydd church, Kent 245 

149. Brixworth church from the north-west - _ - - - 246 

150. Springing of arch at Brixworth ------ 247 

151. Plan of Brixworth church ------- 248 

152. Sectional drawing of the east end at Brixworth - - - 251 

153. Interior of Brixworth, looking west - - - - - 252 

154. Brixworth tower, at level of triple opening - - - - 255 

155. Plan of eastern end of Reculver church - - - - - 256 

156. Roman and Saxon columns - ____-- 259 

157. View of Wing church, Bucks ------ 260 

158. Saxon cathedral at Canterbury ------ 260 

159. Cr^'pts at Ripon and Hexham ------ 265 

160. Plan of Wing church, showing crypt ----- 268 

161. Plan of crypt at Repton ------- 269 



162. Crypt at Sidbury, Devon ------- 271 

163. Pilaster strips, etc., at Gernrode and E.irls Barton - - - 276 

164. Plan of San Nazaro Grande, Milan - _ - . - 282 

165. Plan of early church at St. Denis - - _ - _ _ 283 

166. Altar end of church on Plan of St. Gall - - - - 283 

167. Plan of Hersfeld -_--_-.- 284 

168. Masonry at Stone-by-Faversham ------ 287 

169. Plan of western tower, Deerhurst- - - - _ _ 299 

170. Double aperture at Deerhurst ______ ^oo 

171. Plan of St. Mary, Dover Castle ------ ^07 

172. Plan of Saxon church, Peterborough - - - - - 315 
172 bis. Doorways at Miserden (Saxon) and Hatfield (Norman), - 327 

173. Plan of Bosham, Sussex - - - - - - - 328 

174. Chancel arch at Bosham - - -- - - - 329 

175. Map of Saxon churches ---_--_ ^^^ 

Erratum : p. 86, for 'Fig. 29 ' read 'page 344.' 

Ecclesiastical Architedlure in 

England from the Conversion of the Saxons 

to the Norman Conquest 



In a chapter of the previous volume a place .was found 

for a brief sketch of the ecclesiastical history of the British 

Isles up to the time of the final conversion of the immigrant 

Saxons at the close of the seventh century. At the date of 

the Teutonic descents Romano-British Christianity was an 

established institution in the country. The invaders crushed 

it in those eastern and central parts of England which they 

made their own, but it continued to flourish in the west where 

it was effectively represented by the Church of Wales. Irish 

Christianity, and that of southern Scotland as introduced by 

Ninian, appear to have been in their origin independent of 

the Romano-British Church, and rested rather on Gaul as 

their basis ; but after the accession of strength gained by the 

Church in Wales when the whole Christianity of Roman Britain 

was as it were forced in to it by the heathen impact from the 

east, the Churches of Cornwall, Wales, southern Scotland, and 

Ireland mav be regarded as representing one extensive body 

of Celtic Christianity centering first in Wales and afterwards 

in Ireland, a Christianity that had maintained forms and 

traditions which had come to differ from those prevalent in 

the Romanized west. Saxon Christianity had its definite 

beginnings at the close of the sixth century and its annals 
n A 


for the succeeding hundred years are recounted with some 
detail in the Ecclesiastical History of Bede. 

There are accordingly three classes of Early Christian 
buildings of which the monumental history of these islands 
must take account, (i) the remains, if any exist, of Romano- 
British churches prior to the Saxon invasion ; (2) structures 
built and used by Celtic Christians in the non-Romanized parts 
of the islands ; (3) Saxon churches erected subsequent to the 
conversion of the invaders. Of these the first must be sought 
within the bounds of the Britannic provinces, the second in 
the western and northern portions of Great Britain and more 
especially in Ireland, and the third in Teutonized England, 
that is in the eastern and central portions of the island as far 
north as the Firth of Forth. It has already been made clear 
that Roman, Romano-Gallic, and Celtic influences were all 
brought to bear on early Saxon Christianity, It will be well 
therefore, before embarking on the monumental history of the 
latter, to gain some idea of the manner of planning and build- 
ing practised at the time in the north-western provinces of the 
Roman Empire as well as in the non-Romanized Celtic lands. 

It is proposed to deal here briefly, first, with the materials 
and technique of pagan Roman buildings especially in Britain 
and with the normal plans of Romano-Christian buildings 
throughout the Empire, and, second, with Celtic building 
traditions as they are illustrated for us in Ireland. It should 
be explained that in the following paragraphs account has only 
been taken of those features of Roman and Celtic work which 
have a distinct bearing on that of Saxon times. 

I. The Roman Sources, 

There is no fact connected with ancient Rome that is more 
characteristic than the uniformity of technique in her great 
public works which arose in all the lands under her sway. 
Broadly speaking, the methods of choosing, preparing, and 


employing materials are the same in all the provinces of the 
Empire, on the Euphrates as in Spain, in Africa as in Britain, 
so that the work in our own country is in all essentials exactly 
what we meet with in northern Gaul and on the Rhine and 
the Moselle. 

In the matter of technique Roman work is uniform, and as 
a rule uniformly good. It is however a mistake to suppose 
that genuine Roman construction is invariably marked by great 
accuracy of technique. An examination of the old Roman 
pharos or lighthouse-tower within the castle precincts at Dover 
shows that there may be exceptions to this general rule. The 
monument in question has been greatly knocked about and 
patched at various epochs, but some original portions of it, which 
are still as the Roman workmen left them, are roughly and 
unevenly wrought. At the same time, though haste or 
economy, or a shortcoming in the supply of materials, may at 
times have lowered the standard of execution, yet genuine 
Roman building is always the work of men who have learned 
their business, and possess the advantage of well understood 
technical traditions. A Roman arch may be rudely put to- 
gether, but no one can doubt that the constructor understood 
the principle of cutting and placing voussoirs. 

In the matter of different materials and of the methods 
of their employment, we find represented among the Roman 
remains in England (i) the 'opus quadratum,' or construction 
with large squared stones; (2) the massif of rubble concrete or 
' structura caementicia ' faced with small parallelepiped stones 
with or without bonding courses of brick ; (3) the ' opus testa- 
ceum ' where the fabric or skin of a structure is of brick ; (4) 
the plain wall of irregular stone-work with no special facing or 
technique; and finally (5) the light partition of wood-work and 
plaster. The following brief notes on these materials and methods 
contain some facts of which we shall need to take account. 

As an example of Roman opus quadratum the jamb of a 
Roman gateway still standing at the station of Cilurnum 



(Chesters) on the Northumbrian Wall may be taken as typical 
(Fig. i). Some noble specimens of massive Roman stonework 
have just come to light at Castlecary near Falkirk. 

In some of our Roman structures very large stones may 
be found employed as footing, or to form the upright jambs 

- -.4 

Fig, I. — Jamb of Roman Gateway at Chesters, on the North Tyne. 

and sills or lintels of doorways, and these vertical and horizontal 
pieces are occasionally mortised into each other after the fashion 
shown in Fig. 2, where the slabs lining the doorway measure 
more than 6 ft. in height by 2 ft. 6 in. in width, and have a 
ridge cut out along their top to fit into a corresponding groove 
in the lintel, which has now disappeared. 

One peculiarity of the large squared stones used in this class 
of work is their tooling. They are often scored by the pick 
with diagonal indentations that sometimes cross each other so 
as to form diamonds, or with more deeply marked semi- 



circular grooves forming sometimes a sort of pattern. See 
Fig. 3. It is advisable to take note of this Roman treatment 
of the surfaces of stones, as it enables Roman stones to be 
recognized when re-used, as is so often the case, in Saxon 
wallins. The toolincr on such re-used stones has sometimes 

to b 

been signalized as ' Saxon.' 

. It'.- «M- „■"*«■ n 

J>«~ V -«"» 

1 ^-^ 

Fig. 2. — Roman Doorway at Chesters on the North Tyne, showing 
the jambs mortised into the sill and lintel. 

: The Roman method of construction in rubble concrete is 
represented abundantly in Britain in walls both of a civil and 
a military character. The regular small squared stones, with 
which these are commonly faced, give them a very distinctive 
character. Lines of brickwork composed of two or three flat 
tiles superimposed occur very commonly at intervals of a few 
feet. The mortar in which the small stones or bricks are set is 
often compounded with coarsely pulverized tiles. 

Roman brickwork is less common in the north of Britain 
than in the south, but it is not unknown in the higher latitudes. 


There is evidence of its use on more than one site in Scotland, 
and quite recently it has been found as far north as the Roman 
station at Inchtuthil in Perthshire, where has come to light a 
Roman bath. The north however has always been a stone 
country, and in all the structures connected with the Roman 
Wail between the Tyne and Solway this is the material par 





Fig. 3. — Tooling on Roman Stones in the Northumberland Wall. 

excellence. We find here walling of large squared stones and 
of smaller material very carefully cut and set, but the stations 
on the Tyne also furnish partition walls of rubble stonework 
not specially faced, that resemble walls found in Roman villas 
all over the country. These walls differ from the regularly 
faced ones by their thinness and slightness of technique. They 
are commonly from i foot to 2 feet thick, whereas the faced 
walls, as must necessarily follow from their technique, are far 

In Roman wall construction in squared stone work as a 


general rule no special technique is observable at the quoins, 
which are usually carefully made up in the same material as 
the walling. The peculiar Anglo-Saxon quoining which goes 
by the name of ' long and short work ' is certainly not in 
evidence in Roman buildings in this country. 

The use of plaster is of course abundant. A special kind is 
made with pounded tiles and is red throughout, not merely 
flecked with the red of testacean fragments. It is very hard 
and impervious to water and is used for the lining of bath 
chambers and for floors as well as for the coating of walls 
generally. The term ' opus signinum ' is commonly applied 
to it. The deeply scored tooling on Roman stones already 
referred to was perhaps originally intended to afi-ord a key for 
plaster, though it appears on stones that never seem to have 
been so coated. The wood-and-plaster partition walls men- 
tioned above have left their traces in some domestic buildings 
excavated at Silchester.^ 

The employment of cut stone work for arches, and of rubble 
concrete for vaults, was no doubt as common in this country as 
elsewhere in Romanized lands. The Roman bridges of which 
Bede writes need not all have been arched structures, as they 
may have consisted, like Rochester Bridge in Saxon times, in 
piers of masonry with a superstructure of woodwork,^ but some 
of those in chief use were no doubt of masonry throughout. 
No arch of a Roman bridge survives in this country but portions 
of the stone piers of Roman bridges are still to be seen, as for 
example near ChoUerford on the North Tyne. 

The best existing Roman arch of stonework in the solid opus 
quadratum is the already noticed Newport arch at Lincoln, the 
only complete gate of a Roman city that still survives in situ 
and in use." A small chamber vaulted with large stones roughly 
cut in voussoir fashion still exists at Cilurnum on the North 
Tyne, but Roman vaults were in general far more frequently 
of concrete than of cut stone. Li our own country the existing 

'^ Archaeologia, lvi, 243. "^YoX. i, p. 82. ^V'ol. i, p. 58 f. 


remains of such vaults are very scanty. We possess nothing 
resembling the great vaulted chambers of the Thermae by the 
Musee Cluny at Paris. In Britain the most important interior 
that is known to have been vaulted is a chamber at Uriconium 
to the south of the Basilica, measuring 49 ft. by 33 ft. 
Traces indicate that this was vaulted by six intersecting groined 
vaults, but only their springing is preserved.^ 

For arches not in stone or concrete the use of flat bricks set 
edgeways, often in two rows one outside the other, is common, 
and these bricks are sometimes made in voussoir form thicker 
at one edge than at the one opposite. The alternation of bricks 
and stone voussoirs is common, and the stone used is often tufa 
which the Romans, and after them the Normans, favoured for 
use in arches and vaults on account of its lightness. As 
regards openings it may be noted that doors and windows are 
usually cut straight through the wall in orthodox classical 
fashion. The recessed openings which occur at the so-called 
imperial palace at Trier have not been noticed in England, 
but a remarkable example of an internally splayed Roman 
window occurs at Cilurnum (see Fig. 41, postea, p. 93). 

Sufficient remains exist to show that the Roman cities of 
Britain were supplied with handsome columnar edifices. Among 
these, basilicas seem to be chiefly represented, the remains of 
Roman temples that can be identified being very scanty. 
Bases, portions of shafts, and capitals of columns that were 
as much as twenty to thirty feet in height have been found, as 
at Lincoln, Wroxeter, and Bath. The capitals are sometimes 
debased corinthian but generally of a modified Roman doric, 
in which is apparent a tendency to elaborate the classical 
annulus into a series of mouldings. The bases are attic. 
In two mediaeval churches near the Roman Wall in 
Northumbria Roman monolithic column-shafts are used in the 
nave arcades. The churches are Chollerton on the North Tyne 
and Lanchester in county Durham. There are in all six com- 

1 G. E. Fox, Uriconium, in Archaeological Journal, liv, 14.7. 


plete shafts, seven or eight feet in height and about five feet in 
girth, together with four others now half embedded in the 
walls to form responds. They are of sandstone, and are not 
brought to a finished surface, but are covered with tooling that 
seems partly Roman (Lanchester western respond) and partly 
mediaeval. One of them is shown in Fig. 156, postea, p. 259. 
The outline of the shafts is so irregular that repeated testing 
with the straight edge and measuring tape leaves it doubtful 
whether there was any intention of giving them the classical 
taper and entasis, though there are some indications of these. 

One special class of columns calls for particular notice. 
These are small shafts some three or four feet high with 
attic bases and moulded caps that show distinct marks of 
having received their form by being turned in the lathe. This 
peculiarity is found in pieces ot all sizes, and Mr. G. E. P^ox 
states that 'in every Roman site in Britain where columns, or 
capitals, or bases are found, there is evidence of the lathe being 
used in forming them.'^ Examples in the Leicester museum 
and at Chester exhibit this evidence very clearly. What these 
shafts were used for is not quite clear. There are plenty 
of roughly blocked pillar-like pieces on old Roman sites that 
formed the ' pilae' or supports of the upper floors of hypocausts, 
but the shafts in question are too finely wrought to be them- 
selves pilae of this kind, though they often share with the hypo- 
caust pillars a bellying form. Mr. Fox refers to them as 
* dwarf columns the uses of which it is not easy to define,' and 
states that ' those of small size were certainly occasionally em- 
ployed as the supports of stone tables,' - while ' others of larger 
dimensions, placed on a dwarf wall, upheld the roofs of peri- 
styles in domestic buildings . . . possibly they may have served . . . 
as dividing shafts to large window openings in gables.' ^ 

^ Archaeological Journal, xlvi, 48. 

2 In the Museum at Mainz one may be seen so employed, while at Silchester 
evidence of the same use was found. Sec Archaeologia, Liii, 280. 

'•''Arch. Journ. liv, 170. 


Two specimens of these shafts are given in Figs. 4 and 5. 
One is a little shaft 3 ft. 3 in. high, with cap and base of 
orthodox though debased form, that occurs, re-used, in the 
belfry opening of the Saxon tower at Wickham, Berks ; the 
other, a roughly hewn stump, of about the same size, comes 
from Housesteads on the Northumbrian Wall. 

Shafts similar to these are sometimes introduced on a small 
scale as ornaments on carved stones such as tombstones or 
altars. These details, though they may seem in themselves 

Fig. 4. — Roman Shaft re-used in 
belfry stage of Saxon Church 
Tower, Wickham, Berks. 

Fig. 5. — Roman dwarf Pillar at 
Housesteads on the Northum- 
brian Wall. 

insignificant, become of importance in connection with the 
turned baluster shafts of stone which occur in Saxon work of 
apparently very early date, and their significance will appear as 
we proceed. 

Mosaic pavements, the tesserae of which are formed of native 
stones or small testaceous cubes, are common especially in 
villas. The patterns of Roman mosaic pavements have a 
special interest in that they appear in some cases to have sug- 
gested motives which occur on sculptured stones and other 
forms of Old English ornamental art of Christian use and date. 
For the moment however we are only concerned with what 


belongs to the domain of architecture, and the consideration of 
these patterns must be postponed. 

The foregoing technical notes are from secular or at any- 
rate non-Christian structures. There is in this country one con- 
spicuous instance of Roman construction applied to a Christian 
purpose in the frequently mentioned Early Christian basilican 
church excavated a few years ago at Silchester in Hants. 
Other fragments in existing early churches are claimed as 

Fig. 6. — Plan of Early Christian Basilican Church in the Roman City at 
Silchester, Hants. Copied by permission from Archaeolo^a, liii. 

Roman but their attribution is a matter of uncertainty, whereas 
at Silchester the work is undoubtedly Roman, and the only 
debate that can be raised about it is whether or not it is 

The situation ot the structure in question in relation to the 
Roman city was shown in Vol. i, Fig. 19, p. 146. Fig. 6 gives its 
plan indicated by the lower courses of the walls, which were laid 
bare in the year 1892, when the writer had an opportunity of 
seeing them, but are now again covered in. Their material was 
flint rubble with brick quoins. The plan shows a narthex, a nave 
terminating in an apse, and side aisles ending in spaces marked 
off on either side of the apse, and projecting on the exterior 
beyond the line of the aisle walls. The building, the total 
exterior length of which was 42 ft., was oriented with its apse 


towards the west. The floor was laid with a pavement of 
red tile tesserae about an inch square, but in the centre of 
the apse was a square space in which was a mosaic pattern 
the date of which, from a comparison with other Roman 
mosaics, is estimated in the report in Archaeologia ^ as the 
fourth century a.d. There was no trace of an altar or 
of any seat round the apse or at its central point. In 
spite of the absence ot these, and of any distinctively 
Christian mark such as a cross or monogram, the building 
in its situation, plan, and orientation so exactly corresponds 
to what would be expected in a Christian church of 
the date indicated that it would be pedantic to doubt its 
Christian attribution. There are no definite Christian marks 
in the way of ornaments or symbols in the case of other early 
churches such as Reculver. They are known as churches by 
the continuous ecclesiastical tradition that attaches to them, 
but such a tradition could not exist at Silchester where the 
town passed out of existence before the middle ages began. 
The building in question may have had a wooden table -altar, 
which was the natural form at the period before relics came 
into fashion, and is the form indicated in the mosaic pictures 
of Christian altars of the middle of the fifth century in the 
Baptistry at Ravenna. The seats round the apse, if indeed 
there were space for any, may also have been of wood. 
The western orientation at Silchester suits an early period, 
as the priest would minister standing with the altar between 
himself and the people and in this position would face the 

As bearing on the question of the Christian character of the 
Silchester structure, we may note that in the numerous Early 
Christian churches of North Africa the altars were also evidently 
of wood," and no trace is now left of them save where the floor 
mosaic has preserved indications of the places where once they 

Uiii, 563. 

- Gsell, Les Monuments antiques de PAlgerie, Paris, 1901, 11. 145. 


stood. Thus at Sidi Mabrouk in Algeria there were mosaics 
all over the interior, but in the middle of the apse there v/as a 
rectangular space left plain, thus reversing the arrangement 
at Silchester.^ Similarly, in the Early Christian church at 
Rusguniae in North Africa, of which an account is given in 
the Bulleiin Ardieologique^'- while there is no trace of a per- 
manent altar or of seats round the apse, the floor mosaic, of a 
far more elaborate kind than that at Silchester, exhibits in the 
centre of the apse, just where the altar would stand, a portion 
of a figure of a recumbent Lamb, and in front of it there is an 
inscription which runs across the chord of the arc and begin- 
ning with the letters ARA ends with the letters TERNO, the 
inscription in between being destroyed. Though there is no 
such Christian figure or lettering at Silchester, a fact of which 
the sceptic has made the most, a comparison of these three 
apses cannot fail to suggest that they had all the same 

This little Christian church at Silchester is properly 
described as Roman, and hence it is assumed that its proto- 
types v/ould be found in Rome or in the other cities of the 
peninsula. It may be of advantage however to pause here a 
moment to consider what is the true meaning of this term 
' Roman,' as applied to Early Christian monuments or institu- 
tions. Certain elements of the civilization which we call 
' Roman,' and which we find pervading all the lands of the 
Empire, had in fact their origin and the source of their power 
in the actual city of the seven hills ; such elements were law, 
and the military and administrative system in general, but it 
must not be supposed that Rome herself was the fountain 
head of all the intellectual and religious movements which 
pulsated throughout that vast domain. Christianity, . to take 
the most important of these, was not indigenous at Rome. 
In this case Rome was colonized by the adherents of a religion 

^ ibid, p. 259. 
-* 2 Paris Comit^ des Travaux Historiqucs, etc., 1900, p. 126 f. 


that had its original centres in the East, and Christianity 
radiated from these original centres over the Empire along 
lines that by no means necessarily passed through Rome. 
Hence the Christian forms and institutions, which are some- 
times called ' Roman ' because they are found all over the 
regions of the Empire, were in their birth and development 
independent of the city, and in dealing with them we must 
remember that Rome was only one out of many centres of 
Early Christian life. 

It is to be noted that the Early Christian meeting-house 
appears in substantially the same forms over the whole of the 
lands of the Empire in its eastern as well as its western 
moiety. Whatever was the origin of the type of building 
known as the Christian basilica,^ we meet it in the first age 

^ This is still doubtful, and the suggestion lately made by Dehio and von 
Bezold, which has been adopted in the new Anglo-American Handbook of 
Christian Antiquities (Macmillan, 1 901), to the effect that the atrium of the 
Roman house gave its original form to the church, does not appear a very 
helpful one. The atrium, or central hall, of the house was just that part of it 
least likely to be used for the semi-secret gatherings of the Christians during 
the ages of persecution. It is true as the above writers point out that the 
alae of the Roman atrium might account for the Early Christian transept in 
the form in which we find it in so many of the basilicas of the city 
Rome, but then this is practically confined to the city Rome and 
is by no means a normal basilican feature, while on the other hand the 
apse, which is a normal feature and the most constant of all, has no 
prototype in the atrium. On this interesting and much discussed question 
the writer sees no reason to change the view which he put forward some 
years ago, to the effect that though meetings in private houses affected the 
surroundings and adjuncts of the Early Christian church, the peristyle of the 
house becoming the forecourt of the church, yet the special architectural form 
of the church is not to be sought in any part of the private house. We may 
imagine the small unobtrusive churches of the third century conforming to 
the type shown in the well-known Early Christian sarcophagus of the Lateran, 
where we find indicated a simple oblong interior terminated by an apse at the 
end opposite the entrance. This is a form suitable for what was known as a 
'schola,' which means of course not a 'school' but a lodge-room or place of 
meeting of any brotherhood or company of people who have business to 
transact under the presidency of officials. Such a building would become a 


of ecclesiastical architecture in a large number of different 
provinces, and find it everywhere essentially the same though 
with local modifications that are well worthy of attention. 
The chief provinces of the architectural type may be enumerated 
as (i) Central Syria, (2) Asia Minor, (3) Thrace, Macedonia, 
Greece, (4) Egypt, (5) North Africa, (6) Rome, (7) Italy in 
general, (8) Gaul, Spain, and Britain. The variations from 
one district to another are sometimes marked enough for us to 
speak of distinct local schools, but as a rule they are too 
slight for this to be possible. An extensive enumeration of 
examples in the different regions, with indications of their 
peculiarities, is to be found in the article ' Basilica ' in Kraus's 
Encyclopedia,^ and one particular set of variations, those 
that distinguish the basilicas of the North African province 
from the basilicas of Rome, has lately been discussed by 
M. Gsell in his valuable work on Algerian monuments."- 

Some of the variations concern the presence or absence 
of an atrium or columned court before the church, which 
is common in Syria, in Rome, and in Gaul, but is hardly 
known in North Africa ; or of a narthex or long vestibule 
across the entrance end, which is not an Italian fashion but 

basilica of the normal shape if the side walls were broken through and rows of 
columns substituted to give access to aisles. This very process went on in most 
of our English parish churches where aisles have been added in similar fashion, 
and the same thing may well have been done much earlier. On this hypothesis 
we readily account for the principal features of the church, including the apse, 
and we also account for the fact that the normal Early Christian basilica has no 
galleries over the side-aisles, though this was so common a feature in the pagan 
basilicas of the Roman cities. When once the side-aisles were introduced they 
became universal even in buildings small enough, like Silchester, to be roofed 
in one span, and the reason probably is that the tripartite division of the 
interior was convenient for separating the different sections of the congregation. 

^ Real-Encyklopddie der christlkhen Alterthumer. 

2 Les Monuments antiques de r Alger le. M. Gsell notices no fewer than a 
hundred and thirty-eight Early Christian churches of which remains are to 
be found in the district, and he notes that they show very little variation 
one from the other in plan and arrangement. 


is common in the East and in northern Africa. The altar 
end of the building is variously treated. The church is 
sometimes spread out at the end where the services are 
performed in order to accommodate a larger body ot 
rninistrants or give facilities for ecclesiastical display. Rome 
has here a speciality in the transept, which appears in many 
of the basilicas of the city, but is not found elsewhere till 
it reappears a little later in some ot the early mediaeval 
churches of Gaul. Subsidiary spaces for the purposes ot a 
sacristy are variously arranged in different examples and in 
different groups, and M. Gsell specially notes that the 
flanking chambers on either side of the apse called 
' Prothesis ' ^ and ' Diaconicon ' are rare in Italy ^ but in 
North Africa almost universal. 

It is somewhat remarkable that in several ot the points 
in which Italian basilicas differ from those of North Africa 
and the East, Silchester agrees distinctly with the latter 
and not with the former. Its type is not Roman or 
Italian but rather North African or oriental. At Silchester 
there is a distinct narthex quite on an oriental plan, while 
the apse is flanked by two chambers of which that on the 
north appears to open into the presbytery and so to cor- 
respond to the diaconicon or service-chamber universal in 
North Africa and the East, while the corresponding chamber 
to the south may answer to the prothesis, which in the 
regions just mentioned opens towards the church rather 
than the presbytery. The floor at Silchester was all on one 
level while in the African churches the presbytery is 

^ So called as the place where people deposited, or * put forward,' oblations. 

- The best examples here are the early basilica of Sta. Sinforosa outside Rome, 
and San Spirito, Ravenna. San Giovanni a Porta Latina at Rome might also 
be claimed as an Italian example, but the walls which here divide nave and aisles 
at the altar end date only from the seventeenth century. The spaces at the 
ends of the aisles are only chapels, and the sacristy is in another part altogether. 
See Crescimbeni, Istoriadella chiesa di S. Giov. av. Port. Lat. Roma, 171 6, p. 86. 



generally raised, and this constitutes a difference which does 
not however nullify the remarkable resemblances here 
noticed. These seem to bear out in a striking manner 
what was said in the previous volume about the probability 
that the Christianity of Gaul and Britain was at first 
independent of Rome, and in touch rather with the East, 
and it will be remembered in this connection that the 
language in the chief centres of the Gallic church in 
the second century was Greek and not Latin. It is 

G e 


1 1 

Fig. 7. — Plan of Early Christian Church at Guesseria, North Africa. 
From Gsell, Les Monuments Antiques de F Algerie. (No scale.) 

probable therefore that the little church at Silchester 
was affiliated to this partly oriental Gallic church, which 
shared its eastern connections with that of northern Africa. 
The accompanying plan. Fig. 7, taken from the work of 
M. Gsell, presents as near a parallel to Silchester as could 
well be found. It is of similar proportions but just 
double the size, and like the African churches in general has 
the apse towards the east. The narthex is present in both, 
while the extension of the lateral chambers on each side of the 
apse beyond the line of the aisle walls is another striking point 




of similarity. This extension may seem at first sight like a 
transept but it is essentially different, as the side spaces are 
here subsidiary and do not, like the transept, form an integral 
portion of the open interior of the building. 

II, The Celtic Sources. 

Early Christian architecture in the non-Romanized parts of 
Great Britain and in Ireland is represented by numerous extant 
monuments, that offer a series of successive types affiliated on 
pagan structures and ending with Christian churches of simple 
but matured form such as are found in other parts of Christen- 
dom. Among the Celtic districts of these islands, Cornwall 
and Wales, the Christianity of which goes back without break 
of continuity to Romano-British times, offer disappointingly few 
well-preserved architectural remains of early date. Scotland is 
in this respect far richer, and the existing monuments of the 
class under consideration have been fully illustrated, partly 
from the notes of earlier investigators, in the recently pub- 
lished work by Messrs. M'Gibbon and Ross.^ The Celtic 
work in question is however best represented in Ireland, the 
Early Christian art of which, through the medium of lona, 
largely influenced that of the neighbouring Caledonia. For 
the practical purpose of this chapter it will be sufficient to 
draw examples from the comparatively well preserved structures 
of Ireland, it being understood that work of the same kind 
though more fragmentary occurs in all the other Celtic regions 
above enumerated. 

The early types met with in these regions appear in their 
main features to be of purely native origin. Neither in their 
form nor their general technique do we discern any copying 
of foreign models, until in the later examples the use of lime- 
mortar and the arched window heads and chancel arches 
betoken a borrowing from Roman traditions. The series of 
types may be briefly described as follows. 

^ The Ecclesiastical Architecture of Scotland, Edinburgh, 1896-7. 


It begins with the circular so-called ' bee-hive ' cell of dry, 
that is uncemented, stonework. Such a structure can be seen 
in the drawing of Skellig Michael given in the previous volume.^ 
It is there employed for Christian purposes, but the form can 
be followed far back into pagan times. Among the most 
imposing of Irish monuments are certain great sepulchral 
tumuli beside the river Boyne, at Newgrange and at Dowth, 
recognized as burial places of ancient Irish kings, and belonging 
undoubtedly to the pre-Christian period. Access to the 
interiors of these immense mounds is gained by passages lined 
and roofed with huge upright and horizontal slabs of unhewn 
stone, and these end in chambers, circular in plan, and in 
section resembling a bee-hive, wholly constructed of unworked 
stones put together without cement and forming domical vaults 
by the system known as encorbelment. The structures forcibly 
remind the traveller of the tombs or so-called treasuries at 
Mycenae and other sites in Greece. At the ' tomb of Atreus ' 
at Mycenae there is the same tumulus of earth with a domed 
sepulchral chamber of stone in the heart of it, the stones being 
all laid flat but each ring being a little smaller than the one 
below it, so that the structure gradually grows to a point as it 
rises. In Greece the stones are carefully cut to shape and 
their lower edges which overhang have been chiselled away 
so that the concave surface is everywhere smooth. In Ireland 
the stones are rough but of vast size, and the primaeval great- 
ness in the effect of the monuments lends them a majesty which 
the more finished classical examples cannot surpass. 

Chambers or cells of this same bee-hive shape but of lesser 
size and constructed of smaller pieces, such as would be picked 
up on stony sites, occur in the open in many districts of 
Ireland and on the western sea-board of Britain generally. 
Many no doubt have once been covered by mounds of earth 
and served a sepulchral purpose, but others were certainly 
dwellings and must have been open from the first. They 

1 Vol. I, p. 152. 


are frequently found within the areas of the great stone 
forts that are such remarkable features of the west of Ireland 
and of Wales. These forts are often of a most imposing 
character that is largely helped by the nature of their 
surroundings. That known as Dun i^ngus, on the largest 
of the Aran islands in Galway Bay, of enormous strength, is on 
the very edge of a sheer cliff washed three hundred feet below 
by the Atiantic waves. The Irish appellation for these forts 
is ' cashel,' a word borrowed from the Latin ' castrum,' 
' castellum,' and they come into the story of Christian 
architecture because their deserted precincts were some- 
times occupied, like Roman 
enclosures in England, by 
Early Christian monastic com- 

The dry-stone ramparts 
which are of immense thick- 
ness, as well as the circular 
walls of the bee-hive cells, are 
pierced by doorways the con- 
struction of which is of interest 
for our theme. In the example 
illustrated in Fig. 8 the jambs 
of the doorway are composed 
of huge upright stones like those in the passages of the tumuli, 
and the lintel of a horizontal stone, above which in other 
examples we sometimes find another in order to relieve the first 
from pressure. One notable peculiarity of such doors is the 
fact that for the purpose of reducing the obligatory length of the 
lintel-stones the jambs of the doorway are made to converge, so 
that the aperture is smaller above than below. This feature occurs 
also in the monumental doorways to the Mycenaean tombs. 

It happened more than once that the interior of a stone 
fort or cashel of this kind was taken possession of by a 
Christian community. Of this we have literary as well as 

Fig. 8. — Doorway in an old Irish fort. 


monumental evidence. A well-known example is Innism.urray, 
off the coast of Sligo, where an enceinte, almost certainly 
of a non-Christian origin, encloses Christian buildings of an 
early type. Among these are some bee-hive cells exactly 
like those already referred to, but these might conceivably 
be survivals from a pre-Christian time when the place was 
a secular fort. On Skellig Michael on the other hand (see 
the illustrations, vol. i, pages 152 and 198) the structures are, 
one and all, certainly Christian. The massive wall of enclosure 
forms a Christian cashel, and is constructed just like the 
pagan stone forts, while some of the bee-hive cells are stamped 
as Christian by a cross formed by white quartz stones over 
the doorway. These however show a divergence from the 
simpler form hitherto discussed in the fact that, though 
circular externally, they are rectangular, with sharp or rounded 
corners, in the interior. 

These Christian bee-hive cells on Skellig Michael were 
dwellings not chapels, as is shown by the fact that there are 
openings in the roof for the egress of smoke, but there 
is some reason to think that the circular form may have 
been used for oratories. In a passage already referred to 
about the primitive Irish monastery (vol. i, page 197) we 
have the dimensions given of certain ecclesiastical structures 
which are ascribed to St. Patrick himself. As only one 
dimension is in each case given, it has been argued that 
the buildings were round, so that the diameter gave the size. 
If this do not seem quite convincing, we have the further 
significant fact that we are told several times of ' ecclesiae quad- 
ratae,' and this adjective would hardly have been used had 
not the round form been at the time familiar. If the round 
oratory were known in the earliest Irish period, it is conceivable 
that it would develop into the normal rectangular form after 
the manner indicated in these transitional cells on Skellig 

The development was inevitable tor the reason that an 


interior round in plan cannot be oriented and offers no 
natural place for the necessary altar. Orientation, though 
the Church of the earliest times generally and the Church 
at Rome throughout was indifferent to it, was made much 
of in later days by universal Christendom. The disadvantage 
from this point of view of the round or polygonal interior 
was felt on the Continent, in connection with monumental 
buildings of the type of San Vitale at Ravenna or San 
Lorenzo at Milan. These have been oriented in a somewhat 
forced fashion by building out a special altar-house or in 
some way emphasizing one diameter of the circle ; but the 
contradiction between the architectural character of the 
building and the Christian scheme of arrangement for an 
interior is still apparent. 

The Irish, who in common with the Christians of the other 
parts of these islands seem to have made a special point of 
orientation, would be met by the same difficulty when they 
tried to use the traditional circular plan for purposes of 
Christian worship. The cells on Skellig Michael at any 
rate exhibit certain stages in the transition from the round 
to the rectangular form that are interesting to trace. The 
largest of the cells of the first character on the rock is 
oblong not round in external plan, and possesses a window as 
well as a door, so that it may conceivably have been an oratory, 
in which case the cross marked with white quartz stones over 
its portal would have a special significance. This may give 
us an existing example of a type prior to the evolution of 
the ecclesia quadrata. The latter is represented at Skellig 
Michael in a very primitive-looking and interesting monu- 
ment. This is a small rectangular oratory, oriented consider- 
ably north of east, that stands somewhat apart from the other 
structures just above a projecting corner of the cashel wall and 
on the edge of an almost precipitous descent to the sea. Its 
general form is shown in Fig. 9. It measures only about 8 ft. 
in internal length and is entirely constructed of slab-like 

(,'// IJBS 






















pieces of slate, untouched by the tool. These are placed 
from the ground upwards in the system of encorbelment, 
so that all the sides (except on the end where the door 
comes) converge from the first, and the whole looks like 
a truncated pyramid with convex sides on a rectangular 
base. The door has a flat lintel and sloping jambs that 
are not formed however of single upright stones but are 
built up like the rest of the walling. The external ledge 
on the south side is hardly part of the structure, but was 
evidently arranged to furnish a warm and sheltered seat on 
this retired spot for the lonely and meditative votary. 

A larger building of this type, near Kilmalkedar on the 
mainland of Kerry, shows an east window which has the 
peculiarity that it is splayed both outside and in, the central 
aperture being only 6 in. wide, and is surmounted by a flat 
lintel, while beneath it are the remains of the ancient stone 
altar. The roof was composed in the same fashion by 
encorbelment, but it has given way in the upper part. Per- 
fectly preserved on the other hand, with the exception of 
the coping, is the famous oratory called ' Gallerus ' on the 
Dingle promontory of county Kerry — one of the most 
interesting little buildings in Europe. 'Gallerus' — the 
meaning of the name is not known— is constructed with 
extreme care and skill, and has preserved its walls and 
stone roof intact for, perhaps, 1200 years. The form is 
the same as that of the oratory on the Skellig (Fig. 9) ; 
but the stones in the interior have been, to some extent, 
smoothed by the hammer, and they lie so closely together 
that the surfaces of contact have probably been treated in 
the same way. This refinement, together with the fact that 
the window, in this case only splayed internally, is arched, 
shows that the example is an advanced one of its type. 
Fig. 10 gives the aspect of this interior masonry in the 
south-east corner. No cement of any sort has been used, 
and it is easy to see daylight through the crevices between 


the stones of the walling. These crevices do not however 
let in moisture, for the stones are bedded with a slight down- 
ward slope towards the exterior, so as to throw off the rain. 

Fig. 10. — Interior of the small dry-stone oratory, ' Gallerus/ 

This slope has also a constructional value in counteracting 
the tendency of the stones, as they overlap on the interior, 
to topple inwards. 

The construction of stone roofs by this process of encorbel- 
ment does not concern us in this place for it has only the very 
slightest bearing on the subject of Saxon architecture, but it may 
be noted that the next stage in constructional development 
brings us to buildings like that shown in Fig. 1 1 . Whereas at 
Gallerus the slope of the vault begins practically from the 
ground level, in the example on St. Macdara's island there is an 
upright wall on all four sides, that is kept lov/ on the flanks so 
as to throw the greater part of the height of the building into 
the roof, but continued on the gable ends to the full elevation. 
These gables, it will be discerned, are acutely pointed and this 
is due to the mode of construction of the roof, for it is obvious 



that the steeper the slope the more stable will be the vault in 
encorbelment. The method of roofing gave rise also to very 


Fig. II.- — Oratory on St. Macdara's Island, off Connemara, Ireland. 

curious features in Irish buildings of this class, that at first sight 

have a somewhat classical appearance (Fig. 12). These are 

upright pilaster-like projections 

that appear to east and west at 

the corners of the building, as 

if they were extensions of the 

side walls beyond those of the 

gables. The visitor to Ireland 

who first sees these thinks of 

Roman work, but they are 

purely native, and are clearly 

connected with the system of 

roofing already noticed. This 

becomes evident in those examples where the projections 

are not confined to the vertical walls, but extend up the 

Fig. 12. — Pilaster at the south-west 
corner of ancient Church at 
Dulane, near Kells, Ireland. 



gables to their apex, presenting the curious appearance shown 
in the St. Macdara Oratory. 

A further development of this characteristic Irish system 
of construction brings us to the churches with double stone 
roofs, in which the outer covering is a vault in encorbelment 
of the type already shown, but this is supported beneath, 
or at any rate the interior of the church is covered 

^cale Of feet- 

Fig, 13. — Sections of Irish buildings with double stone roofs, from about the 

seventh to the twelfth centuries. 

below, by an ordinary barrel vault of round or ellip- 
tical section. Fig. 13 shows some sections of the 
principal Irish monuments of this particular character. 
They are in three groups lettered A, B, and C. In the type A 
there are no vertical walls and the slope of the vault begins 
practically from the ground. In B there are low vertical walls 
and a single stone vault in encorbelment surmounting them. 
In the examples marked C the roof is double, there being 
a lower vault with a space forming a chamber over it.^ 

1 The statics of these interesting structures were briefly discussed by the 
writer in The Builder for October, 1897. 


This system is carried on into the twelfth and thirteenth 
centuries, and the classical example of it is Cormac's chapel 
on the Rock of Cashel, a structure in the Norman style. 
With this after development however we have here no concern. 

The structures in question offer problems of some interest 
to the student of the statics of building. Though in most cases 
there has been a certain amount of restoration in recent times, 
these vaults, both upper and under, seem to have remained firm, 
and have not been reconstructed, while in no case has any 
buttressing of the external walls become a necessity. This 
fact reflects no little credit on the ancient Irish mason, who not 
only evolved a novel scheme of construction but carried it 
out with perfect success into practice. 

Our concern at the moment is rather with planning 
than construction. All the buildings previously noted have 
been single-celled oratories, but there exist in almost all parts 
of Ireland churches that possess not only the main rectangular 
chamber, but a smaller chamber to the east of it, also rectangular 
and forming what we should term a chancel. These nave and 
chancel churches form a large class of buildings represented in 
most parts of Ireland, that show a distinct architectural 
advance, though it does not follow that in individual cases 
single-celled oratories may not be later than churches with 
chancels. There are instances in which a single-celled oratory 
has been enlarged by the subsequent addition of a chancel, but 
in the majority of cases nave and chancel are contemporaneous, 
and we may take this to be normal. The buildings in question 
are plain structures, with a nave generally entered through a 
west door, and lighted by primitive window-slits with a square 
or triangular head, or a round one cut out of one or perhaps 
two stones. An archway leads into the small, square-ended 
chancel, that has usually an east window and one other, of the 
same kind as those in the nave. The old stone altar sometimes 
remains, and this is a feature in which Irish country churches 
have an advantage over our own. Ground plans of two 



characteristic examples are given in Fig. 14, Killiney old 
church, near Dublin, and the so-called 'Trinity' church at 
Glendalough, county Wicklow. Both are roofless and in ruins. 

Fig. 14. — Plans of Killiney Old Church near Dublin, and 'Trinity' Church, 


The latter building has at the western end an adjunct, later 
in. date than the main structure, that has no external doorway 
and cannot therefore have served as a porch. As a rule the 
churches of this class have no projecting porch or other adjunct. 
An interior view of the Glendalough example is given in Fig. 15, 
and it will be seen there that it has a chancel arch of mature 
construction the full width of the chancel. It is to be noted 
indeed about all the churches of this class that the chancel 
arch is comparatively wide in relation to the plan as a whole. 
The most notable difference in technique, between these 
nave and chancel churches and the more primitive ones 
already noticed, is that the former are built with mortar, and 



were sometimes plastered. The material is wrought stone, 
but this varies from coarse rubble to finely cut ashlar. 
The construction is sometimes ot large granite blocks very 
carefully fitted, and has a fine megalithic character. No 
special treatment of the quoins is to be observed. On the 
other hand there are about these more advanced structures 
distinct reminiscences of the older traditions. As regards 
construction, some examples even in the twelfth century^ had 

Fig. 15. — Interior view of 'Trinity' Church, Glendalough. 

Stone roofs single or double such as have been described. In 
most cases however the roof was of wood of the ordinary kind, 
though even here the influence of the primitive technique is still 
to be observed in the facts that the side walls are generally low 
while the gables are acutely pointed. The pilasters on the 
gable ends at times also survive as in the so-called ' cathedral ' 
at Glendalough, a building that measures 30 ft. in internal 
width and can never have had a roof of stone. 

In connection with this question of width it might have been 
expected that the old tradition of stone roofing would have 

^ Kilmalkedar church, Fig. 13, B (not the earlier oratory mentioned on 
page 23), and Cormac's chapel, Fig. 13, C"". 



resulted in a tendency to contract internal breadth. We have 
seen that it led to low walls and steep gables. Why did it not 
also result in long and narrow interiors which would have 
facilitated stone vaulting ? As a fact both the single-celled 
interiors and the naves of the more advanced structures are as 
a rule by no means of specially narrow proportions. It is just 
possible that the influence of the original derivation from the 
round form may have been here at work and have kept the two 
dimensions of the rectangle near each other. There is however 
one building, not in Ireland but possibly erected under Irish 
influence, that is curiously elongated in plan. This is that 
single-celled chapel on the west coast of Lancashire at 
Heysham on Morecombe Bay, dedicated to St. Patrick and 
possibly representing an Irish mission station on this accessible 

coast, of which a plan was 
given in vol. i, p. 312. To 
this structure and to the 
elongated plan which it repre- 
sents we must return on a 
subsequent page. Its Irish 
origin is too doubtful for it 
to be discussed in this place. 
In matters of detail we find 
in these more advanced Irish 
buildings sundry distinct traces 
of survival. These are ap- 
parent in the doorways, which 
in their most characteristic 
form are flat-headed and austerely plain. Fig. 16 shows 
a good example. The position of the doorways is almost 
always at the western end, and in this again we see a survival 
of primitive tradition. The southern doorway occurs but very 
rarely. Lord Dunraven stated that he only knew of three 

^ Notes on Irish Architecture, Lond. 1875 ^^^'y ^°^* "» ^^• 

Fig. 16. — Western doorway to the 
Church of Our Lady, Glendalough. 



The windows whether arched, flat-Hnteled, or covered 
with a triangular head, are internally splayed. The double 
splay of the eastern light at Kilmalkedar is quite exceptional. 
Ornamental details are practically non-existent, and this 
plainness of the early buildings of a people who excelled 
in the same epoch in the decorative arts is not a little 

The accompanying illustration. Fig, 17, shows an incised 
cross (a) that occurs on the soffit of the lintel of the door-way 
in Fig, 16, together with a characteristic detail (/>) in the form 
of a projecting stone like a corbel which occurs sometimes 

a be 

Fig. 17. — Ornamental details from old Irish buildings. 

at the base of the gable of churches of the kind and has 
been termed a ' handle stone,' and {c) a specially Irish finial. 

In the nave and chancel church we have arrived at a 
normal type of church plan such as can be paralleled in many 
parts of the Continent, and at forms and processes, such 
as the use of the arch and plastering, that are undoubtedly 
of foreign origin. The general scheme of the two chambers, 
though found outside Ireland, may at the same time be 
in Ireland of purely native origin. \{ the single rectangular 
chambers were adopted for the sake of securing a suitable 
position for the altar, the addition of a second smaller chamber 
of similar form may be due merely to the desire to emphasize 
still further its sacredness. There is no need to assume a 
continental derivation for so simple a scheme. 

There is no object in following any further the development 
of Irish ecclesiastical architecture, for the next stage is marked 


by the introduction of carved ornament ot considerable richness 
in the style known as Irish Romanesque. This work, in which 
a strong Norman influence mingles with native decorative 
feeling, dates not earlier than about iioo, by which time the 
Saxon architecture with which we are at present dealing had 
passed away. At the earlier work it has been necessary to 
glance, because the question of its relation to that of Saxon 
England must necessarily occupy our attention in the sequel. 



We have now glanced at the forms and methods employed 
in building by the Romans and by the non-Romanized 
Celtic peoples, and have studied at Silchester and in Ireland 
some of the Early Christian structures in which these forms 
and methods are illustrated. The churches erected in 
England subsequent to the Christianizing of the invaders 
would be expected to show considerable traces of the influence 
of these pre-existing traditions, and this influence would 
naturally be strongest at the beginning, when the Teutonic 
converts were for the first time essaying the novel art of 
church construction in brick or stone. 

A caution given in the last chapter may usefully here be 
repeated. It must be clearly understood that in the use of 
the word ' Roman ' no direct importation from the city 
Rome or even from Italy is asserted or implied. By Roman 
forms are meant the forms in use all over the Roman 
Empire, in Gaul and North Africa as well as in the Italian 
peninsula. There has been a general tendency in discussion 
upon early Saxon forms to turn at once to Italy for proto- 
types, but the reader will be asked to agree that as a fact 
there is very little need to bring Italy into the argument. 
We can find what we want in the way of prototypes much nearer 
home. The form of Wilfrid's crypts, for example, has been 

II c 


connected with that of the ' cubicula ' of the Roman Catacombs, 
but the truth is that such barrel-vaulted chambers are common 
enough in Gaul and Germany, where we find them of various 
ages, from the early examples, probably of the sixth century, 
that have been lately excavated in the cemetery of St. Matthias 
at Trier, Many of the French and German churches of old 
foundation, such as St. Savinien at Sens, St. Louand by Chinon, 
Werden a.d. Ruhr in Rhenish Prussia, possess such chambers 
as a part of their crypts. What there is really Roman in early 
Saxon work can easily have been derived either from the classical 
monuments of our own country or else from the more 
numerous Roman monuments of Gaul. 

A Saxon architectural style may have been in this way 
originally formed upon Roman and Celtic models. The 
style however possesses a history of some length, and in 
the four hundred and fifty years which this embraces it 
would be natural to expect both internal changes and 
developments and the importation of new elements from 
other lands. 

Records of church building occur throughout the Anglo- 
Saxon period. After the first great missionary age was over, 
we read of events like the rebuilding in the eighth century of 
the church at York,^ with its thirty altars, that must, one would 
think, have given an impetus to church extension all over the 
North. Subsequent to this came the Danish desolation, but 
even after the inroads had begun we hear of local activity 
in church architecture on the part of bishops such as Swithun 
of Winchester, who died in 862 a.d.- The close of the next 
century witnessed a widely-diffused revival encouraged by king 
Edgar and carried out under Dunstan, iEthelwold, and 
Oswald. Cnut, at the beginning of the eleventh century,^ 

^ The description is given in a poem attributed to the famous Alcuin and 
printed in Historians of the Church of York, Rolls Series, No. 71/1, p. 394. 
- W. of Malmesbury, Gesta Pontificum, Rolls Series, No. 52, p. 161. 
2 W. of Malms. Gesta Regum, ad ann. 1017. 


built and rebuilt numerous churches and monasteries, and 
In the middle years of it one of the laws ascribed to 
Edward the Confessor refers to the fact that there were 
then three or four times as many churches as in the early 
days, in terms which seem to imply that church building was a 
feature of the age just prior to the Conquest.^ 

We can obtain a convenient division of the time from the 
conversion of ^thelberht to the Norman Conquest by making 
an early period end with the first inroad of the Danes and the 
ravage of Northumbria, a middle period cover the epoch of the 
Danish wars, and a late one begin with the reign of Edgar. 
These three periods would correspond roughly to the years 
600-800, 800-950, 950-1066. 

Any attempt to distribute the general body of the existing 
Saxon monuments among these periods must be deferred till 
the monuments themselves have been passed in review, but a 
notice of the foreign influences which are most likely to have 
been operative at each epoch will suitably here find a place. 
The object is not to pre-judge at the outset any of the debated 
questions of origin in Saxon architecture, but merely to indicate 
the conclusions to which the writer has been brought, in order 
that the reader may have these in mind as hypotheses during 
the study of the monuments themselves in the chapters that 

Of the three building epochs just indicated there is no doubt 
that the two really prolific ones were the first and the last, for the 
central or Danish period though not a wholly barren epoch 
certainly cannot have been one of great productiveness. It is 
possible nevertheless that new elements were then introduced 
which had an effect on the architecture of the succeeding era 
when church building was again in vogue. 

In the first epoch, as we have already seen, there was 
a mixture of Roman and Celtic influences modified, but 
only slightly if at all, by native Teutonic traditions. In 
^Thorpe, Ancient Laws, etc., p. 191. 


the second epoch, that of the Viking invasions, we might 
anticipate a Scandinavian influence, but it is a question whether 
we can point to any architectural feature for which such a 
derivation can reasonably be claimed. In decorative art, in 
wood-work both of buildings and of ships, and in the sphere of 
manners and customs, the influence of the Northmen may be 
detected or at any rate discussed. The Vikings were in no 
case however stone builders, and cannot have imported any 
new features ready-made into the practice of the architectural 
art in stone, in which they had no native traditions. Any 
influence they exercised would be indirect and might take two 
forms. First, the new social conditions produced in England 
by the Danish invasions might lead to the modification of 
building fashions and the evolution of special forms. This 
was certainly the case in Ireland, where there seems now no 
question that the famous round towers were towers of refuge 
necessitated by the dangers to which monastic communities 
were exposed in the days of Viking ravage, and it has been 
suggested that the pre-Conquest church towers so common in 
the north eastern districts of England may have been due to a 
similar cause. Second, the Vikings, though not themselves 
stone builders, may have familiarized England with timber 
forms which were afterwards introduced into later Saxon stone 

The question of the influence of timber work on Saxon 
architecture has already coine before us. There are those, the 
late J. T. Irvine was one, who have attached great importance 
to this influence and have used it to explain not a few of the 
features of pre-Conquest building. The examples however 
to which they appeal are not the earlier structures, such as St. 
Martin, Canterbury ; Brixworth, or Escomb in Durham, but 
those that are acknowledged to be of comparatively late date 
such as the towers of Earls Barton and Barnack, Northampton- 
shire. These towers exhibit the narrow projecting vertical 
strips of stone work which at first sight seem closely to 


resemble the uprights of the half-timber framed buildings that 
were noticed in the opening chapter of the first volume. The 
buildings which show these vertical strips are of the Danish or 
the later epoch, and it is conceivable that the Danish settlers in 
our eastern counties may have so stimulated the timber 
architecture of the old country that it gave birth to these some- 
what remarkable features of later Saxon stone buildings. 

This will accordingly be a convenient place in which to 
inquire what were the probable forms of the timber work em- 
ployed alike by the original Saxon settlers, the Celtic missioners, 
and the later intruders from the Scandinavian North. It is 
unlikely that there would be marked differences among these. 
The expression ' mos Scottorum ' is used more than once^ for 
the wood building of the Irish, but this need not imply that 
there was any special kind of construction in wood peculiar to 
the Celtic tribes ; we have no ground for assuming that the 
early Teutonic conquerors of Britain brought with them any 
particular technique different from those practised by timber 
constructors all over the world, nor that the methods in wood- 
work of the Danes or Norwegians of the ninth century 
represented any marked technical advance on the those of the 
Jutes or Frisians of the fifth. There are essentially three forms 
of wood technique, represented respectively by the wattled hut, 
the block house, and the structure of framed timber work. 
The first is merely an application to building purposes of the 
very primitive processes of mat or basket weaving, one of the 
very earliest of the crafts, and cannot in its very nature attain 
to monumental dignity. The two last rise to the rank ot 
architectural processes. There is this difference between the 
block house technique and that which employs the frame and 
filling. The former is simpler and more natural and is used 
where the material is abundant, whereas the latter is scientific 
and effects a saving in material. In the block house method the 
tree trunks or squared logs are placed close together either in a 
^ e.g. Bede, Historia Ecclesiastica, iii, 25. 


vertical or a horizontal position and form a compact wall like that 
of a Swiss hay chalet or Canadian backwoodsman's hut. In the 
framed system the beams are disposed at intervals, and some 
commoner and less resisting material is used to fill in the 
spaces between them. 

It is of course this latter style of work that is claimed as 
having originated those features of Saxon stone architecture just 
referred to. It was so familiar in our own country in later 
mediaeval times that its existence at remoter epochs has been 
assumed as a matter of course. It is however contended here 
that this assumption is not justified, and that pre-Conquest 
timber work was more probably in the compact block house 
style. The available arguments on both sides may be thus sum- 
marized. Framed timber work was familiar to the Romans, 
who had to work in highly civilized lands where wood was 
no longer abundant. There is evidence that they used it in our 
own country at Silchester ^ and elsewhere. Again, we must 
admit a knowledge of the principles of framed timber work 
on the part alike of the early Saxons and Angles and of the 
later Vikings, for the art of ship-building that they practised in 
their continental homes cannot be carried far without the intro- 
duction of the ribs and planking, which in their mutual rela- 
tions are not unlike the frame and filling of the half-timber 
house, M, Ruprich-Robert indeed, in his work on Norman 
architecture, offers a suggestion (which he does not follow out) 
about the possible influence of the tradition of ship-building 
among the northern sea-rovers on the timber-work of the roofs 
of the later Norman churches,- The ship had however neces- 
sarily to combine lightness with strength, an end secured by 
nailing comparatively thin planks over a framework of strong 
pieces disposed at intervals. Were a house built like a ship it 
would have the uprights inside, and the planks would present 
a uniform outer surface like that of the modern hoarding, 
but very unlike that of the half-timber dwelling. Ship- 

^ Archaeologia, lvi, 2-43 f. "^ V archttectiu e Normande^ Paris, 188-I-, ch, vii. 


building technique, in other words, would not naturally lead to 
the framed timber work of the mediaeval house. In the conti- 
nental homes however both of Teuton and of Norseman, and 
in the England where they settled, timber was at first so plenti- 
ful that they would not trouble to do more than cut the number 
of logs required and fit them together in the simple and solid 
block house fashion. It is in the highest degree improbable 
that they would adopt the economical methods of later mediae- 
val times, when wood was becoming scarcer, and would space 
their uprights at wide distances apart, using wattle-and-clay or 
similar materials for the filling-in. Yet, if we suppose the 
pilaster strips ot eleventh-century Saxon stone churches to be 
imitations of timber-work, we should have to assume that this 
advanced mediaeval technique was already in vogue at least as 
early as the tenth. 

It is natural to refer in this connection to the well-known 
timber churches of Norway, on which Professor Dietrichson has 
published an instructive monograph,^ These Norwegian struc- 
tures, highly interesting though they may be, are not exactly of 
a primitive character. They are basilican in plan, were very often 
apsidal, and in such features as round-arched arcading or cubical 
capitals, exhibit a direct imitation in wood of Romanesque 
stone forms, while no earlier date than the eleventh century is 
claimed for any existing example. They are in fact in many 
of their features rather copies of stone buildings than their 
prototypes. The attempt made by Ruprich-Robert to derive 
from this source some of the Norman architectural forms has 
met with no acceptance.^ 

The construction of the Norwegian wooden churches is 
advanced, and depends, as Professor Dietrichson has shown, 
essentially on the principle of frame and filling. To argue 
from these to the English structures of more primitive times 
would be extremely hazardous. The writer last quoted tries 
to prove a connection between the Norwegian structures and 
^De Norske Stavkirker, Christiania, 1892. ^ V architecture Normande, loc. cit. 


the one example of a Saxon wooden building that still remains 
to us, the nave of the church at Greenstead in Essex, already 
more than once referred to in these pages.^ What now does 
Greenstead offer to us ? 

The church, which is a mile or so from Chipping Ongar 
just beyond the bounds of Epping forest, consists now in a 
modern chancel of normal type, a nave measuring internally 
26 ft. by 17 ft. and a western tower. The last is coated with 
wooden planks and is comparatively modern, the nave is the 
Saxon fabric and the walls of it are composed of upright balks 
of timber, made of trunks of oak trees split down the middle, 
stripped of their bark, and smoothed with the adze on 
their flat faces. They are placed closely side by side with 
the flat faces inwards and the half rounds showing on the 
exterior. The joints between them are covered internally 
with modern strips about two inches wide. The general 
aspect of these wooden walls may be judged from the view 
of the north-west corner given in Fig. 18. What is seen 
now is a reconstruction. Originally the split trunks were let 
into a sill of oak at the bottom and fastened at the top with 
wooden pins to a horizontal plate, but the lower parts of 
the fabric had become rotten through the ground damp, 
and in the year 1848 it was taken to pieces and the 
uprights laid out on the ground for examination. The 
lower portions of them were then cut off and they were 
remounted as they stand at present on an oaken sill upon a 
low wall of brick. 

^ Vol. I, p. 26, 37. In connection with Greenstead reference must be made 
to a document associated with St. Edmundsbury, that is given in Dugdale's 
Monasticon, iii, 139. It is there stated that in the year 1013 the body 
of St. Edmund was conveyed from London to Suffolk and rested for a night 
near Aungre (Ongar) where a chapel was constructed of timber for its 
reception. Apud Aungre hospitabatur ubi in ejus memoria lignea capella 
permanet usque hodie. There can be little question that this is the very 
structure that has come down to us. 



The technique here is the simplest possible. The walls 
are simply strong palisades. The sill and the plate are mere 
adjuncts for fixing purposes and are essentially posterior, 
not prior, to the uprights that represent the main structure. 
There is no sign of that skeleton which in framed work is 
essentially prior to the filling and of heavier section. It is 
necessary to insist on this as Professor Dietrichson is inclined 

Fig. 18. — North-west corner of timber church at Greenstead, Essex. 

to regard the corner pieces, the sills, and the horizontal lintels 
as answering to the framing of the more scientifically con- 
structed Norwegian churches. As a fact there are at Green- 
stead no corner pieces, but the corner is formed of a bole 
just like the others the only difference being that a quarter 
is taken out of its section instead of its being split in half ; 
see Fig. 19, which shows the ground plan of the north-west 

The same compact or block house system is shown in the 
representations on the Baycux Tapestry of the timber 



structures that surmount the moated mounds, several of 
which are figured ^ in the needlework. They seem put 
together just in the manner illustrated at Greenstead and 
furnish an additional argument against the theory that the 
ordinary half-timber work of later mediaeval times goes back 
to the Viking age. 

These arguments, which so far as they go militate against 
the professed derivation of Saxon pilaster strips from wood- 
work, will be strongly reinforced when the real origin and 
history of the features in question come to be investigated. 

Fig. 19. — Plan of north-west corner at Greenstead. Scale, J^ of nature. 

It will be shown in the sequel that the Saxon pilaster strips 
are not sui generis and are not of British origin. Their 
prototypes are to be found abroad, and neither in their 
original home on the Continent nor among ourselves are they 
derived from work in timber. 

On the whole then we must negative the hypothesis that 
either Danish or earlier Saxon timber technique supplied models 
for Saxon stone architecture. 

Putting now aside the idea of direct Danish influence 
exercised through wood technique, we may ask whether any 
other foreign sources suggest themselves for the features of 
Saxon architecture in the middle and later epochs. As in the 
earliest epoch so here also, Italy has been laid under con- 

^ Vol. I, p. 1 10. 


tribution, and a direct Roman origin has been claimed for some 
of the later as well as the earliest of our Saxon forms. The 
hypothesis of an Italian derivation has seemed indeed to present 
itself naturally to the minds of most of the writers who have 
discussed pre-Conquest forms, and it is necessary to inquire 
how far this evident predisposition can be justified. 

Attention has been already directed to the special attraction 
which, from the days of Benedict Biscop and Wilfrid, was 
exercised upon the English by Rome. The intercourse with 
Rome begun in the seventh century was close and long- 
continued, and would reasonably account for the adoption of 
Italian fashions in architecture. The letters of Boniface show 
that in his time there was in some quarters a morbid craving 
for the spiritual exaltation won by a visit to the tombs of the 
Roman saints, and devout women not a few embarked in 
consequence on the long and hazardous pilgrimage.^ The 
English were for centuries among the best customers of the 
Roman purveyors of relics, and the frequency of pilgrimages in 
the later Saxon period is attested by the effort made by Cnut to 
secure a safe passage to and fro through central Europe for 
those bent on the holy mission.'^ And it was not only persons 
of a specially devotional or mystic turn of mind that were 
drawn to Rome. Alfred had an ecclesiastically minded father 
who sent him there as a youth, and later on he accompanied 
thither again his parent iEthelwulf; Cnut however and 
Harold, the two strongest and least ecclesiastically minded men 
of the England of the last half century of Saxendom, also 
visited the sacred sites and secured their relics with as much 
pious devotion as the rest of their fellow-kings and subjects. 

These numerous English pilgrims, all receptively inclined, 
might easily, one would think, have brought back with them 
new architectural ideas that they would put in practice at home. 
When they had deposited their precious burden of relics 

^ Bonifatii £/>. 30, 32, 53, in Mlgnc, Patrol. Curs. Compl. lxxxix, 726 f. 
^Cnut's 'Declaration.' See Licbermann, Gesetxe der Angelsnchsen, \, 276. 


within the altar of some English church perhaps of their own 
advowson, what was more likely than that they would alter or 
enlarge or rebuild the edifice in a style reminiscent of the holy 
sites that had yielded the treasure ? This natural suggestion 
seems to find confirmation in the fact that one characteristic 
Late-Saxon architectural form does occur abundantly in Italy, 
whence at first sight it would seem to be derived. This is the 
double, triple, or multiplex opening divided by one or more 
shafts in the belfry stage of towers, a feature almost universal 
in the Roman Campanili, and one that is familiar, in its simplest 
form of the double opening, in the bell towers of the eastern 
districts of England. This proof of connection has appeared so 
cogent that the Saxon bell tower has come to be very commonly 
regarded as an Italian importation, and we may take as typical 
the remark of Professor Willis ' that the Saxons did imitate 
Roman models is shewn by the very midwall shafts of the 
Saxon windows which are directly copied from those of the 
Roman Campanili.' ^ 

The facts of the case however seem to be that these belfry 
openings, though classical in their origin, were not derived by 
our builders immediately from Italy, but from a much nearer 
region of Europe that had itself borrowed the feature at an 
earlier period from its Italian originators. Familiarity with its 
use in Italy may have influenced our later Saxon builders, but 
they did not themselves import it from thence. It came to us 
from Italy not directly or at an early period, but at a late period 
and through the medium of Germany. The proof of this is to 
be found in the fact that in this country the feature occurs in 
close connection with other features that are not Italian but are 
German, while it can be shown that there is an equal a priori 
likelihood of a derivation from Germany as from Italy. This 
brings us to the general question of German influence on Anglo- 
Saxon England, upon which some remarks were ofl^sred in the 
previous volume.^ 
1 Architectural History of Canterbury Cathedral, p. 3c. ^Yo\. i, p. 215 f. 


The marked similarity in certain special features between 
later Saxon buildings and those of a large region of Germany is 
a fact that no unprejudiced observer will deny. If it be not so 
generally recognized as are the supposed Italian affinities, this 
is due to the fact that local phases of architecture in Germany 
are not so familiar to us as those of the more attractive and 
more freely visited peninsula. In the chapters which follow 
opportunity will be taken to point out in detail the resemblances 
here spoken of, and in this place it will be sufficient merely to 
glance in a general way, first at the political and social ties that 
linked together Anglo-Saxon England and the Germany of 
Carolingian and later times, and next at the characteristics of 
German architecture in those features of it that seem to bring 
it into special connection with our own. 

In the latest Anglo-Saxon period the time of Harold, bishop 
Stubbs notices the connections in poHtical and religious 
matters which then existed between England and Germany.^ 
In the earliest or pre-Danish epoch of Anglo-Saxon culture a 
similar comment was made from the archaeological side by the 
late Albert Hartshorne in his well-known work on English glass 
vessels.^ He there points out as a somewhat unexpected fact, 
that whereas at the close of the seventh century Benedict 
Biscop, when he needed the aid of workers in glass, sent for 
them to Gaul,2 in the next century Cuthbert abbot of Jarrow 
in a similar case transmitted his request for the aid of experts in 
glass working to his countryman Lul at Mainz upon the Rhine.* 

1 ' The intercourse of England with Germany was close at this time. The 
Emperor had married a daughter of Cnut, half-sister of the King : the 
Athelings, Edmund and Edward, had married nieces of the Emperor. . . . 
German Clerks were at the head of the Wessex Church.' The Foundation 
of Waltham Abbey, Oxford, 186 1, p. ix. 

^Old English Glasses, Lond. 1897, p. 113. 

3 Bcde, Historia Abbatum, in Baedae Opera Historica, cd. Plummer, Oxford, 
1894, 1. 

* Migne, xcvi. 839. 


The fact is really the key to a good deal of the artistic 
history of the period. It signalizes the concentration of the 
culture of the Carolingian dominions in the Rhineland provinces 
which formed the eastern wing of the Prankish realm, and 
emphasizes the already established connection of Anglo- 
Saxondom with that region. The empire of the Franks, 
which was enlarged and consolidated by the chieftains who 
preceded Charles the Great, extended from Spain to Saxony 
and Bavaria, but from the first it tended to fall into two halves, 
the earlier Neustria and Austrasia prefiguring the later France 
and Germany. In the time of Charles the Great there were 
two chief centres of culture one in each division of the empire, 
Tours upon the Loire and Charles' own favoured seat at 
Aachen near the Rhine. The English came into contact with 
both west and east, for Alcuin was settled at Tours and planted 
in its school his own Northumbrian learning, while Willebrord 
and Boniface wrouo^ht for the conversion of Frisia and central 
Germany. We have seen that these missioners were accom- 
panied and followed by a large number of Anglo-Saxon fellow- 
workers^ so that a special connection was at once set up between 
our own country and the regions beyond the Rhine. The 
intercourse was not only religious but political. The Carolin- 
gian court was the recognized refuge for English political 
exiles" and at the close of the eighth century the famous 
Ecgberht of Wessex had stayed with Charles the Great in this 
capacity for more than a decade. A little earlier than this 
there was intercourse between Charles and the most powerful 
of the English local kings, Offa of Mercia, and Ecgberht's son 
and successor, j^.thelwulf the father of Alfred, took for his 
second wife a daughter of Charles the Bald, while in a later 
generation grand-daughters of Alfred and sisters of king ^thel- 
stan wedded the Carling Charles the Simple, and the Emperor in 
Germany, Otto the Great of Saxony, and established a tradition 

^ Vol. I, p. 216. 

2 J. R. Green, Short History of the English People, ch. i, §14. 


of alliance which was maintained as we have just seen to the 

In the domain of culture and of the arts the character of the 
connection between England and Germany changed from what 
it had been in the earlier Carolingian period. At the epoch of 
Northumbrian greatness the Prankish world was receptive, but 
the first half of the eighth century brought about an alteration 
in the mutual relations of the two regions. Northumbria as 
we have already noted had declined even before the first Danish 
keel put to land at Lindisfarne, while the distinction of the 
court of Charles the Great was as great in learning and in the 
arts as in laws and arms. If the glory of the Carolingian age was 
dimmed under the weak successors of Charles who opposed so 
feeble a resistance to the V^ikings, the rule of the Ottos of 
Saxony again ushered in a flourishing period of European 
culture and art. 

In one sense the Old Saxony of the tenth century was like the 
New Saxony, in its insular home, of the seventh. In our own 
Kent and Northumbria at the earlier epoch, a Christian culture 
partly fashioned on Roman models had been established in 
regions that had been almost wholly de-Romanized ; in the con- 
tinental Saxony of the post-Carolingian period, a region that had 
never formed part of the Empire, and had received its Chris- 
tianity as an importation at a comparatively recent date, became 
the centre of an activity in learning and art that the Germans 
claim as the first truly native expression of their national genius. 
The religious life, and the learning which depended on the 
religious life, of the regions in question were as we have seen 
largely owed to the emissaries from Anglo-Saxon England, but 
so soon as the line of Henry the Fowler was established in its 
imperial state, this quarter of Europe, saved by its inland 
position from the ravages of the Vikings, developed a culture 
which rose above the level that was anywhere else attained. It 
is one of the outstanding facts of the architecture of the central 
mediaeval period that among existing buildings of importance 


that exhibit the Romanesque style in its developed form, the first 
in point of date is the convent church of Gernrode in Saxony. 
The minster of Charles the Great at Aachen which in the books 
begins the long series of Romanesque monuments, is not Roman- 
esque but rather Early Christian in style. Gernrode, which 
dates substantially from the last half of the tenth century is 
genuine Romanesque, and there is scarce another monument 
of its class the date of which is so early and so assured. 

The whole architecture of this great region north-east of the 
Rhine, won for Christianity under the Carlings, assumes a special 
character that we shall do well to note. The region embraces 
what we know as Thuringia, Saxony, Westphalia, Rhenish 
Prussia, and the provinces of the lower Rhine ; and the 
architecture of it differs from that of the other parts of 
the vast Carolingian empire. The old political distinction 
between Neustria and Austrasia is here reproduced, and the 
Romanesque of the western or Neustrian part ot the 
empire develops on lines distinct from that of the Aus- 
trasian regions extending eastwards beyond the Rhine. 
The first is represented centrally by the architecture of Nor- 
mandy, and Norman forms differ in many marked charac- 
teristics from those of Westphalia or of Saxony. 

This fact lies at the foundation of any systematic treatment 
of the later Anglo-Saxon buildings. In several of their most 
characteristic features these only reproduce what is common in 
the Trans-Rhenane provinces, and though Anglo-Saxon build- 
ings have other very distinct features of their own which give 
the style independence, yet they have so much in common with 
German ones that we shall probably be right to reckon our 
own country, in the century before the Norman Conquest, an 
autonomous province of Austrasian architecture. Later Anglo- 
Saxon architecture, it should be clearly understood, has no 
special affinity with Norman, but on the contrary, till the 
fusion of the two realms at the Conquest, it represented a 
quite distinct architectural tradition. At the time of the fusion 


there do appear forms uniting Saxon and Norman charac- 
teristics, but until the Norman element definitely makes its 
appearance the two styles have very little in common, and for 
this reason it is as a rule comparatively easy to distinguish a 
Saxon from a Norman structure. 

The task now before us is first to indicate these character- 
istics in which Austrasian Romanesque differs from Neustrian, 
and next to give the reasons for including later Anglo-Saxon 
architecture in the former province ; only those points which are 
of real significance for the purpose in hand will be taken into 
account. It will be convenient to adopt the following division, 
and test the principle here laid down with reference to (i) 
technique (2) distribution of the parts of a building (3) treat- 
ment of wall-surfaces (4) openings (5) details. 

(i) The architecture of the Prankish realms had an existence 
before the time of Charles the Great though it is impossible to 
date its early monuments with any assurance. There is a class 
of these monuments however that have remarkable and early- 
looking peculiarities of technical treatment. They appear on 
the whole to be pre-Carolingian, and they exhibit a modifica- 
tion of Roman technique in a direction corresponding to the 
tendencies of Merovingian times. They are built on a Roman 
method with core and facing, and exhibit the so-called ' petit 
appareil ' of Gallo-Roman monuments in which the facing 
stones are of the small square Roman shape and are often 
seamed with lines of brick. The curious St. Jean at Poitiers, 
and the church of Vieux-Pont-en-Auge near Mezidon in Nor- 
mandy are good examples.^ The special peculiarity of the class 
of buildings in question is the diversifying of this facing by a 
studied mosaic in which geometrical patterns are formed by zig- 
zags, hexagons, herring-bone work, stars, etc. The best known 

^ These buildings are now generally dated later, see Enlart, Manuel 
d'' Archedogie Fran^aise, Paris, 1902, i, 155 f. They arc however so strikingly 
unlike Carolingian work in Germany that tlie earlier date seems in the mean- 
time equally probable. 

11 D 


examples of this work are at Cravant near Chinon on the Vienne ; 
St. Christophe, Suevres ; Savenieres, and especially the so-called 
' Clara Thurm ' at Cologne, almost the only bit of pre-Roman- 
esque building now visible in that ancient city. The work is 
nothing but an extension of the Roman fashion of facing with 
' opus reticulatum,' herring-bone work, and the like, and the 
elaboration of it corresponds to the sumptuousness in personal 
attire and accoutrements that characterized the Merovingian 
princes and nobles.^ 

This mosaic-like distribution of facing stones is of course 
in less pretentious forms familiar in Norman architecture, 
and the south-west corner of the infirmary cloister at West- 
minster shows an early example of it in English Norman 
building. It hardly occurs however in the Austrasian pro- 
vince, save in the already quoted example of the Clara Thurm 
at Cologne and one other small Rhineland building of peculiar 
historical importance. 

This is the entrance gatehouse to the cloister of Lorsch 
between Worms and the Odenwald. It is figured in all the 
architectural books, but has recently been made the subject 
of a careful monograph by R. Adamy ^ which throws a 
welcome light upon its origin and character. The arguments 
there urged appear satisfactorily to fix the date of the notable 
little monument about thirty years earlier than Charles the 
Great's minster at Aachen, or about 764 to 774 a.d. It was 
built under the auspices of a brother of Chrodegang, bishop 
of Metz, the author of the canonical rule, and with the 
help of monks transferred to Lorsch from a convent near 
Metz, from the neighbourhood of which city part of the 
material of the structure was conveyed. It is therefore not 
an architectural product of its own locality, but really belongs 
to the region west of the Rhine, and will thus fall into line 
with the other mosaic-faced structures of the Neustrian province. 

1 Vol. I, p. 233. 

2 Die fr'dnkische Thorhalle und Klosterkirche zu Lorsch, Darmstadt, 1891. 



The monument is one that we shall have to examine more 
narrowly, but for the moment all that needs to be said about 
it is that it is no exception to the rule that ornamental facing 
treatment of stonework belongs to the western provinces of 
the Prankish empire. 

It is noteworthy that neither in the minster at Aachen nor 
in any other genuine Carolingian structure, such as the basilica 
of Eginhard at Michelstadt in the Odenwald, do we find this 
ornamental facing, nor is it a feature in Saxon or Westphalian 

Fig. 20. — Front and side view of the western end of the Minster at Aachen, 
according to the restoration by C. Rhoen. 

architecture. It is a non-German peculiarity, and it does not 
occur in Anglo-Saxon architecture save in one or two 
examples of herring-bone work of which the most remarkable 
is found in the interior of Diddlebury church, Shropshire. 

The presence or absence of classical core-and-facing 
technique forms then the first point of difi^erence between 
the western and eastern provinces of the original Prankish 
realm. We pass now to the second point. 

(2) The distribution of the parts of a building. An 
indication of one or two points of special significance is all 
that can here be attempted. The influence of the central 



or domed church in the development of European architecture 
is a well-established fact, and few examples of this type have 
exercised such influence more strongly than the minster of 
Charles the Great at Aachen. For the purpose in hand we 
are not concerned with the main structure, the general plan 
of which is that of a central octagon with concentric aisle and 

gallery above, but rather with the 
imposing western forebuilding 
which contains the entrance portal 
and serves other purposes presently 
to be explained. This has been 
altered in its upper portions in 
later times but the original form 
can be recovered, and it seems 
to have presented the appearance 
shown in Figs. 20 and 21 which 
are founded on the restoration 
published some years ago by 
C. Rhoen of Aachen. 
The front view, Fig. 20, shows us that the facade is almost 
entirely occupied by a wide and very lofty but shallow niche, 
the lower part of which is pierced by a spacious archway 
giving access to a vestibule at the end of which is the square- 
headed portal of the church, closed by great bronze doors. 
On the upper story on the level of the gallery round the 
main octagon there is a space corresponding to the vestibule 
below, and this apparently opened to the facade through a 
large window, divided up, like some late Roman and 
Byzantine windows, by columns in two pairs separated by 
an architrave carried by three small arches. Towards the 
octagon also this space was screened off from the gallery by 
columns. Two stair turrets, irregularly semicircular in plan, 
flanking the forebuilding give access to this upper space, and 
the stairs were carried still higher till they ended on the level 
of a chamber forming the third story of the building. This 

Fig. 21. — Section of western end 
of the Minster at Aachen. 


chamber was used for the bells, to one of which is attached 
a story containing so valuable an indication on the subject 
of campanology, that it is added in a footnote.^ On this 
western front of Aachen it will be necessary to linger 

In the introduction to an elaborate study of the Carolingian 
buildings at Werden a.d. Ruhr in Rhenish Prussia a recent 
author, W. Effmann, writes as follows about the historical 
importance of the treatment of the western ends of early 
Romanesque churches. ' Western forebuildings ' — for so the 
writer's compendious term ' Westwerke ' may be translated — 
' in their purpose their origin and their various forms have 
hardly yet been touched by the investigator. They are often 
confused with western choirs and western transepts, and 
their influence on the development of the tower and the 
facade has never been properly recognized.'- 

Such systematic study of the western ends of Romanesque 
churches as is here spoken of must start from Charles' minster 
at Aachen. This contains in itself the germ of many of those 
features the varied treatment of which supplies interest to the 
history of the Romanesque facade. If the flanking stair turrets 
be brought into greater prominence and the central portion 

^Charles the Great gave a commission to a bronze founder at Aachen for 
a great bell and supplied him with a quantity of silver to mix with the 
baser metals in his alloy. The fraudulent craftsman abstracted the 
silver and used a cheaper substitute. When the bell was cast and hung 
in the bell chamber it refused to move until the founder came beneath 
it to try his skill with the ropes. Thereupon down came the huge clapper 
of the bell upon the guilty head beneath and its weight was such that 
it crushed him into a shapeless mass. The legend is recorded by the monk 
of St. Gall about the year 884 so that bells of large size are established, 
at least in the most important centre of European art, as early as the 
time of Charles the Great, or at any rate that of his biographer. See 
Pertz, Mon. Germ. Hist. Script. 11, 744. 

2 W. Effmann, Die KaroHngisih-Ottonischefi Bauten zu Werden., Strassburg, 
1899, p. vi. 


between them correspondingly reduced, the result is the twin- 
towered facade with vestibule and western gallery between the 
lateral masses, that is so common in advanced Romanesque in 
all lands north of the Alps. This form of facade composition 
belongs to northern Romanesque in general, but there are other 
forms of composition suggested equally by Aachen that are 
specifically German. If the central forebuilding be emphasized 
it will grow into imposing though at times somewhat clumsy 
masses like those which are reared at the western ends of 
churches such as the Liebfrauenkirche at Maestricht ; 
St. Patroclus, Soest ; or Paderborn. The half-round stair 
turrets, less important in proportion to the centre than at 
Aachen, are in most buildings of this class still retained, but 
they may on the other hand shrink and disappear so that the 
central portion stands out alone as a single western tower. 

It is a well-known fact that single western towers with or 
without the lateral adjuncts are characteristic of the archi- 
tecture of Westphalia and of the provinces of the lower Rhine, 
but are comparatively rare in other districts of Romanesque 
architecture. That these single western towers of the regions 
named are directly derived from the western forebuilding 
at Aachen is a debateable point, but there is no question 
that a modification of Aachen would without difficulty produce 
them, and that Aachen, by far the most conspicuous European 
building of the early Romanesque period, is a presumable 
source of architectural influence for the eastern part of the 
Carolingian realms. 

Not only the western tower but the western choir also, 
another characteristic German feature, may be found in germ 
at Aachen. In the original form of the building the eastern 
choir and the western forebuilding, both on the ground level 
and the gallery floor, exactly correspond. If there is one 
peculiarity more than another which separates the architecture 
of the eastern or German from that of the western or 
French province it is this correspondence, so common in 


Germany, between the entrance and the altar ends of churches. 
The form that the correspondence generally takes in Germany 
is that of a repetition in a western apse of the plan and 
arrangement of the apse at the east, and Aachen especially 
on the gallery floor seems to prefigure the combination. 

What has now been said is sufficient to show that the 
disposition of the parts at the entrance ends of churches 
marks a difference between the architecture of the eastern 
and that of the western province of the old Prankish empire. 
The difference is not an absolute one ; it must not be sup- 
posed, that is to say, that all German facades are different 
from French. The typical Romanesque elevation where twin 
lateral towers flank a lower central portion, is as much at 
home in Germany as in the other architectural districts north 
of the Alps. Indeed the twin-towered facade at Corvey in 
Westphalia is one of the earliest examples of the type that 
can be named. The similar facade at St. Castor, Coblenz, 
dates in its lower part from the end of the ninth century. 
Gernrode, the original scheme of which embraced a front 
of the kind flanked by round towers, was planned about 960, 
St. Pantaleon at Cologne with its square lateral towers a few 
years later, and these early examples prepare the way for 
the use of the same scheme in numberless churches of the 
Rhineland and farther Germany. 

The point is that in Germany at all periods of Romanesque 
we find the single western tower claiming its place as a 
feature of monumental structures side by side with the twin- 
towered scheme, and in the Rhineland capital the noble 
Apostles church and St. Mauritius reared their single western 
towers in the same city that showed the flanked fa^^ades 
of St. Pantaleon and of the ancient cathedral. In some 
German examples, as at Maursmiinster, near Strassburg, the 
flanking towers are retained but a central western tower 
makes its appearance between and overtopping them, while 
in some regions, as in Westphalia and a large part of 


Rhenish Prussia,^ the latter feature suppresses and ultimately 
supersedes the double arrangement, till it comes entirely to 
dominate the composition of this end of the building. 

Save in England alone we do not find this treatment of 
western ends in vogue in any of the other districts of 
Romanesque architecture. Single western towers do occur in 
France, and Dehio and von Bezold, who remark on their 
imposing character, consider that they had originally a defensive 
intent. Examples are to be found at St. Germain des Pres 
at Paris, at Poissy near the capital, and in the regions to the 
west (St. Savin near Poitiers) and the south (Notre Dame des 
Doms, Avignon). They are uncommon however in the northern 
districts with which we are most concerned. In Normandy for 
example the single western tower is in the Romanesque period 
a great rarity,- though one occurs at Notre Dame d'Esquay 
in Calvados. There now exist western towers attached to 
numerous churches in the Duchy, but these are nearly all of 
much later date when the influence of English fashions was 
operative.^ As a general rule in the case of large churches, 
south of the Alps the towerless Early Christian facade,* north 
of it the twin-towered facade, is the prevailing form, but in 
Germany and England the single western tower claims equal 
consideration. This fact is clear, but the significance and 
historical explanation of the fact are problematical. An archi- 
tectural connection between England and Germany is certainly 
suggested, and this would agree with our general reading of 
the historical relations which bound together the two regions. 

^ A survey of the monuments of this region is given in the work edited 
by Paul Clemen, ^z> Kunstdenkm'dler der Rheinprovinz, Diisseldorf, 1 89 1, etc., 
a model publication of its kind. 

2* En Normandie,' writes Ruprich Robert about the 'clochers,' 'nous n'en 
voyons qu'exceptionnellement sur les portes d'entree des nefs.' V Arxh. Norm. 
p. 97. 

3 See de Caumont, Statistique Monumentale du Calvados, Paris, 1846, etc. 

*The twin-towered facades in the peninsula occur mostly in the south 
of it and in Sicily, where they are accounted for by Norman influence. 


The general considerations already adduced would lead to the 
hypothesis that the scheme travelled from Germany to England, 
but it must be noted that there are peculiarities about some 
of our English western towers that appear so far as we can 
see to be original. In point of actual date, again, some 
of our pre-Conquest western towers are earlier than the 
existing single western towers of Westphalia and other 
parts of Germany. It has been suggested above that the 
germ of the single-towered facade may be discerned in the 
western forebuilding at Aachen. As a fact, though Aachen 
in its main feature, the central plan, was immediately influential 
in churches such as Germigny les Pres by the Loire, we do 
not find monumental evidence of any immediate influence 
exercised by its western front. This does not however 
invalidate the hypothesis under consideration. The influence 
of Aachen was continuous, and the west choir of Essen in 
Rhenish Prussia at the end of the tenth century, and 
Ottmarsheim in Elsass at the middle of the eleventh, are 
both modelled on the famous octagon. In like manner the 
western forebuilding at Aachen may have served as prototype 
for the cathedrals of Paderborn and Minden of the early 
part of the eleventh century as well as for the Apostles 
church at Cologne and the very numerous later examples 
in the regions already indicated. 

It is always possible that examples intermediate between 
Aachen, and, say, Minden, at one time existed but have now 
disappeared. In any case, as the object at the moment is 
not to discuss theories, but rather to place facts and hypotheses 
before the reader as materials for future use, this subject may 
now be left. 

(3) Passing then to the subject of the treatment of wall 
surfaces, we have the striking difi^erence between the provinces 
which it is our object to contrast, that the buttress, a marked 
feature of Norman Romanesque, hardly occurs in the Roman- 
esque of Germany, where its place is taken by the so-called 


' Lisene,' a feature with some superficial resemblance to a 
buttress, but differing therefrom in that it is a decorative ' 
rather than a constructive adjunct. Aachen it is true posseses 
buttresses, for these occur in pairs at the angles of the central 
octagon the corners of which they are designed to strengthen 
against the thrust of the domical vault. It is however noticed 
by Dehio that after this the buttress almost disappears from 
German Romanesque though it is found abundantly in the 
western province.^ 

The German Lisene differs from the buttress in the following 
characteristics. The buttress may as in Early Norman buildings 
be of slight projection, but in that case it has substantial width 
and represents a real addition to the strength of the masonry. 
The Lisene is also of slight projection but is at the same time 
narrow so that it hardly increases the lateral stability of the 
wall. While buttresses, especially in Norman work, appear 
first of all as strengthening the corners of buildings, and when 
they are distributed along the wall have generally some rela- 
tion to structure and inner arrangement, the Lisenen are 
disposed along the wall-surface in closer juxtaposition than 
would be the case with buttresses, and are out of relation to the 
internal construction. 

It has been noticed about these German Lisenen, as about 
our Anglo-Saxon pilaster strips, that they look like the uprights 
of half-timber work. We are fortunate however in being able 
to trace back the history of the feature in German buildings till 
we find it originating not in any form of wood-construction but 
in the classical pilaster that is so familiar a feature in later 
Roman architecture. 

To establish this we have only to turn to the small structure 
at Lorsch near Worms already referred to. The restored view 
Fig. 22 2 shows the upper story of the little building diver- 

^ Kirchliche Baukuust, I, 154. 

2 This illustration, from Fergusson's History 0/ Architecture, has been kindly 
lent by the publisher. 



sified with upright pilasters joined above by those straight-sided 
arches, that are also a feature for which a timber origin has been 
claimed. The pilasters are fluted and possess attic bases and 


•\ I 


Fig, 2 2. — Fa9ade of Carolingian Gate-house at Lorsch, near Worms. 

debased ionic capitals, while the pieces forming the straight- 
sided arches are moulded with classical profiles. The rest of 
the features of this remarkable little structure bear out the 
impression of its curious classicism in detail. The attic bases 
and composite capitals of the half 
columns on the ground story and 
the imposts of the three main 
arches (Figs, 23-4), the acanthus on 
the string course between the two 
stories, the cornice with its dentels, 
are all quite orthodox in design 
and in many instances are carried 
out with no little knowledge and 
refinement. R. Adamy's demon- 
stration exhibits the building as the first and most characteristic 
outcome of the classical Renaissance which was in fashion 
at the court of Charles the Great. The mosaic-like facing 
we have already recognized as a development from late 

Figs. 23 and 24. — Base of half- 
columns and impost of 
main arches, Lorsch. 


Roman wall-work rather than as a Renaissance, but the 
details, which are much purer than any in general use at 
the time, show a deliberate return to the models ot genuine 
antiquity. The pilasters here are undoubtedly Roman, and 
if such pilasters be the origin of the Lisenen the latter will 
also have a classical tradition at their back. 

The German Lisenen however, in their normal aspect, as 
slender strips of plain stonework starting from a plinth without 
separate bases and joined at the top by shallow arcades of small 
wall arches, are so unlike these fluted ionic pilasters at Lorsch 
that some may doubt whether there be any real connection 
between the two. At Gernrode for examole such Lisenen 
occur, and as Gernrode, which lies at the northern base of the 
Hartz mountains, is in a wood-building country and one where 
no Roman traditions existed, its Lisenen might reasonably be 
claimed as independent creations or as derivations from timber 
construction. As a fact however there are many intermediate 
examples, in which debased classical bases and caps attached to 
the long slender strips place the ultimate derivation of these 
from classical pilasters beyond reasonable doubt. St. Castor at 
Coblenz for example exhibits such bases and caps to the strips 
on the lower stage of its twin western towers of the ninth 
century, and the west front at Trier with the east end of Mainz, 
both of the eleventh century, present the same features. 

The German Lisenen therefore descend from an original 
source in the classical pilaster. That the Anglo-Saxon pilaster 
strips in our own country are derived from the German 
Lisenen, is the view to which the reader will be asked to give 
his assent. 

(4) In the matter of openings we have to consider (a) the 
double or multiplex opening, commonly used in the belfry 
stages of towers to affbrd a passage for sound, and (Z*) the 
small window pierced for the admission of light. In both of 
these the forms used in Germany are different from those 
found in the French provinces, and in both cases Anglo- 


Saxon work resembles German, while the Normans are true 
to their own local traditions of Neustria. 

The subdivision of a large round-headed opening by vertical 
piers occurs in some of the Roman Thermae, and the device 
is displayed in very prominent fashion at Aachen, where 
columns in two stories divide the large openings from the 
gallery to the central octagon, and the window in the west 
front, etc., while a single fluted pilaster, to which we shall 
have to return, divides a small incidental opening that can be 
seen in Fig. 20 in front of the northern flanking turret of the 
western facade. From Aachen the arrangement was adopted 
in the tenth century in the parts at Essen which are modelled 
on the Carolingian minster, and it is claimed for Werden a.d. 
Ruhr at an earlier date.^ Whatever be the earliest history 
of the feature, it becomes normal in German Romanesque 
of the eleventh and twelfth centuries, and almost all the 
numerous church towers of that region and period show 
these subdivided apertures. 

It is not the mere subdivision of the single opening that 
concerns us, for this is certainly not confined to the eastern 
province, and occurs commonly in belfries in Normandy 
and other Romanesque districts. The point of importance 
is the special contrivance by which the arrangement is 
engineered and which can be judged of from the accom- 
panying drawing. Fig. 25. This method diff^ers from that 
adopted, e.g. in Normandy, but agrees in the main with 
that which we shall find represented in the double openings 
of our Anglo-Saxon buildings. The characteristic difference 
resides in the fact that in the German example. Fig. 25, as 
well as in Anglo-Saxon work, the whole thickness of the 
wall is dealt with at once, whereas the Normans recess the 
opening so as to decrease in step-fashion the thickness of 
the wall until only the middle portion remains to be sup- 
ported by the central shaft. 

^ EfFmann, Die karolingisch-ottonischen Bauten zu Werdettf p. 24.2. 



It may be worth while to pause here for a moment to 
note the various devices by which in different provinces of 
Christian architecture the wall between two arched openings 
is sustained by bringing the weight down upon a central 
shaft or shafts. The earliest method in point of date is (i) 
to double the shafts, placing one behind the other so that 

it 'h hi <c^<^> 


Fig. 25. — Double opening with mid-wall shaft in the west front of the 
Cathedral at Trier, of the eleventh century. 

their two capitals support the whole extent of the load. 
This occurs in Early Christian work in Italy, as at the 
Baptistry at Nocera previously noticed,^ and is also in 
common use in cloisters. We shall find an instance of its 
employment in Saxon architecture. Italian also is (2) the 
method of corbelling out the capital of a single central 
shaft till its abacus corresponds in length to the thickness 

1 Vol. I, p. 348. 


of the wall. This occurs commonly in the belfry openings 
of Italian campanili. The same method is in vogue in 
German Romanesque, and it is found in one or two examples, 
Sompting, Sussex ; Bolam, Northumberland ; Jarrow, Durham, 
that date before and about the time of the Conquest in 
England. The procedure most generally adopted in our 
pre-Conquest architecture is a modification of this. Instead 
of the capital of the central or ' mid-wall ' shaft being itself 
corbelled out, this shaft is (3) made to support a stone 
slab, or ' through-stone,' which is long enough to take the 
whole thickness of the wall. This through-stone is not a 
capital, for the shaft often has its own capital of the normal 
square plan under the slab, though it sometimes sustains 
the slab without the interposition of any capital. 

A different method is (4) recessing. This which occurs 
in advanced Romanesque generally was specially favoured 
by the Normans. In the recessed opening as just explained, 
the thickness of the wall is brought down by successive steps 
till it is reduced to a width correspondent to the abacus of 
an ordinary capital. When this recessing occurs in openings 
in English buildings they may safely be ascribed to Norman or 
later date, as the method was adopted by Saxon masons only 
occasionally and for archways. In some English belfries which 
have been assigned to a pre-Conquest date, but are really 
Early Norman, there is a curious compromise. Recessing 
is not adopted, and the whole width of the aperture is 
spanned by a single arch passing through the whole thick- 
ness of the wall. An apparent division is however made by 
inserting a sort of frontispiece in the form of a shaft carry- 
ing two subsidiary arches on the external face of the wall. 
This is done in the magnificent Early Norman tower at 
Eaton Bishop, Herefordshire, and it occurs also at Burwash, 
Sussex ; Tugby, Leicestershire, and Wendens Ambo, Essex 
— all structures for which a Saxon origin has been too hastily 
claimed. The compromise shows that the Norman masons 



did not understand the secret of the structural division of 
the unrecessed opening. Fig. 26 shows some examples of 
these diverse methods of treatment. 





Fig. 26. — Methods of dividing a double opening. 

I. Corbel cap on mid-wall shaft, Sta. Pudentiana, Rome. [Italy, 9th century.] 
II. Do., S. Apollinare in Classe, Ravenna. [Italy, 6th or 7th century.] 

III. Moulded through-stone over cap of mid-wall shaft, Church of Hoven, in 

Rhenish Prussia. [Germany, nth century.] 

IV. Corbel cap on mid-wall shaft, Jarrow, Durham. [England, nth century,] 
V. Do., west front of Trier. [Germany, nth century.] 

VI. Do., Sompting Church, Sussex. [England, nth century.] 
VII. Division apparent only. [English-Norman, nth century.] 
VIII. Recessing. [Norman, 12th century.] 

On the historical relations of these methods it may be 
safely said that the corbel cap (Fig. 25) originated in Italy 
where we find it in the numerous towers of comparatively 
early date at Rome and Ravenna, and that it spread from 
Italy to Germany. That Italy exercised an influence in this 
matter over the Rhine provinces is only what we should 

Many doubtful theories have been put forward in archi- 
tectural books about the influence of one Romanesque district 


upon another, but the connection of northern Italy and the 
Rhineland is not to be gainsaid. Lisenen in quite a German 
form occur on the facade of San Zeno at Verona of the beginning 
of the twelfth century, while the resemblance between the decor- 
ative arcading so common in North Italy and Tuscany, as 
on the west fronts of the cathedrals of Parma, Lucca, or 
Pisa, and the dwarf galleries in the upper stages of German 
Romanesque buildings, is too close to allow of a doubt that 
there is historical connection between the regions. The use 
of columns indicates that the influence passed from Italy to 
Germany and not in the reverse direction, for Italy is the 
land of the column, and the presumption is that any 
tradition of the employment of the column would have 
there its origin. It is quite possible that this same Italian 
connection accounts for the extensive use of the column in 
general, as an architectural feature, in Germany as compared 
with France. In Germany, far more frequently than else- 
where in the north, we find the column utilized in nave 
arcades, either in place of, or alternately with, the pier, and 
it is to be noticed also that the German columns so employed 
often exhibit in distinct and even exaggerated shapes the 
classical entasis and taper. 

This particular form of the subdivided opening originates 
therefore in Italy and was conveyed from there to the Rhine- 
land, and it will be made apparent in the sequel that Germany 
transmitted it at a later day to Anglo-Saxon England. 

For the other form of opening of which there is here question 
it is not easy to fix the place of origin, but it is specially 
characteristic of Germany and of pre-Conquest Britain, This 
is the so-called double-splayed window, in which the aper- 
ture for light is at or near the centre of the thickness of 
the wall, and the jambs are sloped away towards the outer 

The Early Christian basilicas at Rome and Ravenna, and 
wherever else they arc sufficiently preserved, possess large 



windows cut through walls that in Italy at any rate are 
comparatively thin. Hence they are either driven in classical 
fashion straight through the walls, or else are very slightly 
splayed in the interior. Later on in early mediaeval times 
the apertures for light were considerably narrowed, and as 
the walls were as a rule correspondingly thickened, some 
kind of splaying to assist the diffusion of light was rendered 
necessary. This was accomplished in two ways, the most 
common method being to locate the aperture in the outer 
face of the wall and splay the jambs and sill towards the 
interior. This method is exemplified in French work gene- 
rally, and is universal in the Norman Duchy, It is used 
also in a Carolingian work of great historical importance, 
the church erected by the famous Eginhard early in the 
ninth century at Michelstadt in the Odenwald in Germany. 
Here the church is lighted by round-headed and by circular 
openings all widely splayed on the interior, and for the 
small windows in churches like Michelstadt of a modest 
size, we may take this form to have been at the time 
normal in all the Carolingian realms. 

About this epoch however, the alternative arrangement 
of the double splay makes its appearance. The lower stage 
of the rotunda at Fulda of 820 a.d. possesses it, and if Pro- 
fessor Adler is right the eastern end of the interesting 
little church at Niederzell, in Reichenau, near Constanz, 
can show still earlier specimens.^ Part of the crypt at the 
eastern end of Werden a.d. Ruhr, of the ninth century, 
exhibits the feature. A series of more or less dateable 
examples, such as one in the north flanking turret of the 
west front of St. Pantaleon, Cologne, of about 980 a.d., and 
one now blocked in the north wall of Bishop Meinwerk's 
Bartholomaus Kapelle to the north of the cathedral at 
Paderborn of the early eleventh century, bring us to the 
central Romanesque period when this feature becomes very 
^ Adler, Baugeschichtliche Forschungen, Berlin, 1870, 1, 9. 


common in the smaller openings of German churches. An 
example from the west front of Trier of the middle of the 
eleventh century is shown in plan in Fig 27. 

In this double-splayed window accordingly we can see 
another peculiarity of the eastern province, for Norman archi- 
tecture is innocent of it, while by its occurrence in the pre- 
Conquest work of our own country it provides another point of 
attachment by which we can associate with Germany our later 
Anglo-Saxon building. 


Fig. 27, — Plan of double-splayed Fig. 28. — So-called 'mushroom' 
window, west front at Trier. capital, from Werden a. d. 

Ruhr, Rhenish Prussia. 

(5) In the matter of details, the only point to which atten- 
tion needs to be specially called is the characteristic forms of 
capital used in the two provinces. 

Carolingian architecture either employed ancient Roman 
caps as in most instances at Aachen, or else imitations of 
antique corinthian or composite capitals executed sometimes 
with no little care. Some of the best of these are to be found 
on the little Vorhalle at Lorsch. The same -imitation of 
classical models continued, and in the vestibule ati Corvey of 
the ninth century, the western building at Werden of the 
beginning, and Essen of the end, of the tenth, and the Bar- 
tholomaus Kapelle at Paderborn of the early part of the 
eleventh, we find elaborate but more rudely executed reproduc- 
tions of the same models. Side by side with jthese classical 


imitations there occur in the tenth century at Quedlinburg ^ and 
at Werden one or two examples of a form of cap for which 
Dehio has invented the name ' Pilz ' or ' mushroom ' capital, 
and which has the profile shown in Fig. 28.2 

From the eleventh century onwards the place of honour 
among German capitals is assumed by the so-called cubical cap, 
produced by the intersection of a cube and a sphere of 
diameter equal to its diagonal, which becomes normal in the 
buildings of the eleventh and twelfth centuries. The origin 
and history of the form are still matters for discussion, and 
some have claimed for it a Byzantine origin,^ while others see 
in it a direct derivation from woodwork.* We are only con- 
cerned here with the fact that, either in its purer geometric 
forms, or else enriched and modified in somewhat fantastic 
fashions, it is characteristic of German Romanesque, but at 
any rate till the twelfth century is but sparingly found in the 
western province. The characteristic Early Norman cap is 
either debased Ionic or else, as at Jumieges and West- 
minster, consists in a parallelepiped block rudely chamfered 
off to fit the top of the shaft. The so-called cushion cap, 
familiar in Norman work especially in this country, is not an 
Early Norman form and will hardly be found before the twelfth 

In England the most common pre-Conquest cap may be 
described as a cubical cap not well understood. It is of the 
cubical type, but the orthodox relation between the cube and 
the sphere is very commonly neglected and abnormal shapes 
result, some of which appear to be sui generis while for others a 
very close parallel may be found in some of the irregular 

^ Wiperti-Krypta and crypt of the Schloss-KIrche. 

2 This is worth figuring because Dehio has suggested for it an Anglo-Saxon 
origin {K. Baukunst, i, 194). The present writer knows however of no such 
capitals in our own country. 

^e.g. Adler, Forschungen, i, 11. 

* e.g. Humann in the Bonner JahrbiJcker, Heft 89, 1889, p. 192. 


German cubical caps just referred to. In their adoption of the 
cubical type the Anglo-Saxon builders seem again to be assert- 
ing their dependence on Austrasian traditions. 

The question with which this chapter opened can now be 
answered by eliminating any hypothesis of direct Danish in- 
fluence, but postulating an architectural connection with 
Germany which must have had considerable importance in 
relation to our native work. It was said above that Anglo- 
Saxon architecture, at any rate in its later phase, represents an 
autonomous province of Austrasian Romanesque. This means 
that in the grouping of the early Romanesque local styles, 
which were gradually evolved from the age of Charles of 
Aachen onwards, Anglo-Saxon work belongs to the German 
rather than to the French connection. It is in many of its 
characteristics directly opposed to the Norman work which was 
destined to supersede it, and in many too it is closely allied to 
that of Germany. It has at the same time its own individual 
features, some of which are inherited from its earlier phases in 
the first period of the age of conversion, while others it seems 
to have developed as it progressed ; and these features it 
employs to the last side by side with those for which foreign 
prototypes, or at any rate foreign parallels, can be found. 
The debt of our pre-Conquest builders to the lands across the 
North Sea may be freely acknowledged, while at the same time 
full justice is done to the substantial amount of originality and 
boldness in our native productions. 




The previous volume contained some notice of the conditions 
under which were erected the monuments, ecclesiastical and 
secular, of the older England before the Norman Conquest. 
There were then mentioned military works in the form of 
town enceintes and entrenchments ; the mansions of kings 
and nobles and the humble dwellings of the burgher or 
villein ; churches of various kinds from the bishop's 
cathedral and the minster of the greater abbey to the 
minutest field church or chapel ; and finally subsidiary struc- 
tures of ecclesiastical use attached to monastic or canonical 

Of all these monuments of various classes the only ones 
that are still effectively represented are the churches. Saxon 
earthworks exist but have no distinctive character and are 
in no way architectural. Some of the town enceintes were 
in masonry, and it is possible that some of the stones now 
visible in the extant ramparts of Exeter or Chester or 
Porchester may have been laid by Saxon hands. There are 
however in these cases no architectural details or technical 
peculiarities to afford ground for identification. Two portions 
of existing castellated structures have been specially singled 
out as Saxon, and it will be well to make it clear that in 



both cases the work in question is Early Norman. One is 
a part of Corfe Castle on the outer enceinte to the west of 
the keep, where is a piece of walling forming one side 
of a former hall or chapel, that is obviously of different 
character from the rest of the structures on the hill. This 
however possesses all the characteristics of Norman work 
and has no shadow of claim to be considered Saxon, The 
other portion of masonry is at Tamworth and forms the 
facing of the embankment of access to the earthen mound 
on which stands the keep of 
Norman and later times. 

The work at Tamworth is a 
particularly good example of 
the kind of masonry known as 
herring-bone work, which we 
have already established as Nor- 
man rather than Anglo-Saxon. 
As the moated mound in general 
is not to be regarded any longer 
as of pre-Conquest date,^ this 
masonry connected with such a 
mound takes its place naturally as 
a Norman production. Fig. 29 shows the manner of it ; the 
characteristic horizontal courses between the bands of herring- 
bone work should be noted. 

Anglo-Saxon domestic structures in so far as they were 
of wood have not survived. The manor house of the period 
may however have been in part at least of stone, and the 
picture of Harold's aula at Bosham in the Bayeux Tapestry 
may be quoted as evidence of this.^ There is no reason 
why portions of pre-Conquest manor houses may not still 
exist embedded, as was the case with Deerhurst Chapel,^ in 
later mediaeval structures, and investigation may yet bring 
some of these to light. 

^ Sec Vol. I, p. 105 f. 2 ibid^ p, 104.. 8 ibid, p. 332. 

Fig. 29. — Early Norman herring- 
bone work, part of the facing 
of an embankment at Tam- 
worth Castle. The stones are 
about I ft. in length. 


In the case of ecclesiastical monuments, though little or 
nothing in the way of subsidiary structures has been 
preserved, a very considerable number of churches are 
represented by extant fragments. Among these are included 
specimens of most of the classes into which as we have seen ^ 
Anglo-Saxon churches can be divided. There are bishops' 
churches, churches of the greater abbeys, town churches, country 
oratories and chapels. The great majority of the buildings 
are however what we should now call by the term ' parish 
churches ' though a certain proportion of these were originally, 
or at one time, monastic. It should be understood at the 
outset that these extant monuments are nearly all buildings 
of the second order of importance, and it must not be assumed 
that they give an adequate idea of the achievements of the 
Anglo-Saxon architect as a whole. To estimate the archi- 
tecture of the period aright we have to take into account 
the literary notices of structures that have now perished, 
some of which were evidentally of far greater artistic pre- 
tension than any of the extant monuments. An estimate of 
Anglo-Saxon architecture on this broader basis will form the 
subject of a later section of this volume. For the moment 
we have to deal with the extant monuments alone. 

Any list of such monuments will embrace a few examples that 
are still more or less completely Saxon, but the majority of the 
items must consist of little more than remains and indications 
that have had the good fortune to escape the ravages of time 
and the zeal of successive generations of builders. In refer- 
ence to these the question may be asked whether it is really 
worth while to catalogue a number of fragments of old 
masonry accidentally preserved, simply on the ground that 
they belong to a specially early period of our architectural 
history ? 

It may be admitted that if this work be done merely in 
the spirit of the collector, it may easily degenerate into 

1 Vol. I, Ch. in. 


something like a ' fad.' To inventory and label so many 
hundred specimens of Saxon masonry as if they were postage 
stamps or beetles is not the proper way to deal with them. 
They have a human and historical as well as an architectural 
value, and this is not to be measured by the number of stones 
that make them up. A few cubic feet of walling are sufficient 
to establish for us on the spot a Saxon village church of 
stone, and this, with all its fittings and surroundings, its 
porch, its altar, its graveyard, was the centre of the social 
movements of that rural community that has remained till 
quite modern times the unit of the national life. It is a 
monumental link between ourselves and the older Britain 
of a millennium ago, and a point round which the patriotic 
imagination may fitly love to play. And further, these 
same few stones, when taken with other better preserved 
examples, may call up before our minds a building that in 
plan and technique may present striking and original features 
and furnish material for a new chapter, or at any rate an 
interesting foot-note, on the architectural history of the 
middle ages. 

From this point of view nothing is really too small to 
notice. Fragments of moulding or carving or masonry, 
insignificant in themselves, may be like the one or two bones 
of the extinct animal, from which the palaeontologist can 
restore the whole organism. Such fragments moreover may 
supply chronological information of essential value, and may 
aflf'ord a means of correcting impressions derived from the 
general appearance and plan of a building. These last are in 
the Saxon period apt to mislead. Thus, for example, as a 
basilican church Wing, Bucks, seems naturally to take its 
place beside the seventh-century basilicas at Brixworth and 
Reculver, but as a fact, even apart from the advanced form 
of its crypt. Wing exhibits details that compel us to place it 
comparatively late. Bradford -on -Avon appears in general 
character a singularly early church, but when we observe its 


double-splayed windows, reckon up its pilaster strips, and 
note the curious resemblance of its external arcading to that in 
the interior of the very late Saxon church at Dunham Magna 
in Norfolk, we begin to distrust the impression of great 
antiquity. On the other hand, the general aspect of the 
porch at Monkwearmouth in Durham would suggest a 
Romanesque rather than an Early Christian origin, but a 
careful interrogation of the various details and ornaments 
leads to the conviction that the work is in reality of the 
early date assigned to it. 

No apology therefore is needed for including mere frag- 
ments among the monuments we have to consider. So far 
as a Saxon character can be assigned with what the writer 
believes reasonable certainty to the monuments in question, they 
are all indicated on the map of Saxon Churches, p. 344.^ This 
represents a personal examination of some three hundred and 
fifty examples that have been signalized as showing signs of 
Saxon origin. Other examples no doubt exist that have 
come under the notice of local observers, though they are 
not yet generally known, and these would repay investiga- 
tion. If this investigation, however, were carried out so 
completely as to cover every visible piece of Saxon masonry 
in all the British counties, the result would still not be a 
final one. There is a possibility that at any moment the 
stripping of plaster from a church wall of uncertain date 
might reveal unsuspected evidence of antiquity in the masonry 
below. It is a recognized fact that in a large number of 
cases the clearstory walls of aisled churches are of earlier 
date than the arcades which were cut through them in the 
Norman or later periods, and a good many of these are 
doubtless survivals from before the Conquest. 

It will not have escaped the notice of observers that we 
find at times an absolutely certain Saxon doorway or window, 

1 An explanatory note is added with the map and index list at the end 
of this volume, where the criteria relied on are indicated. 


as at Somerford Keynes, Wilts, ^ in a wall the masonry of 
which would not in itself have struck the eye as peculiar. 
In the absence of any such definite feature a wall that is 
really Saxon may pass unnoticed, and there may be very many 
such pieces of walling up and down the country. Hence the 
following treatment of the subject can only claim to be pro- 
visional. Fresh facts may come to light that would tend to 
modify the conclusions here reached. These are however 
based on a sufficiently large body of data for them to be 
offered with some confidence to the reader. 

Questions may arise about (i) the number (2) the geogra- 
phical distribution of the monuments, as well as about (3) 
the criteria relied on to establish their Saxon character. 

(i) The notices of the village church in general quoted 
in a previous volume conveyed the impression that Saxon 
England was, in proportion to its population, well supplied 
with churches, and some have gone so far as to say that 
the village church was almost as common a feature in rural 
England before the Norman Conquest as in the days of 
Elizabeth or George the Third. 

There are sufficient incidental references to churches in 
legal and other documents of the early mediaeval period to 
bear out this surmise. Numerous churches are mentioned in 
Saxon land charters and wills and in Domesday, though 
there is no attempt to give a list of them, to discriminate 
their different architectural forms, or to indicate which 
were of stone and which of wood. A Domesday editor 
has remarked that ' to refer to Domesday as in any way 
giving us correct information as to the number of churches is 
useless.'- Domesday notices of churches, as we have already 
seen,'* are apparently fortuitous, and vary for no assignable 
reason in the different counties. Whatever the menlion of 
a church in Domesday may imply, the silence of the 

^ See postca, Fig. 53, p. 102. 2 Domesday for Wiltshire, Lond. 1865, p. Ixvi. 
3 Vol. I, p. 334 f. 


Commissioners is clearly no evidence against the existence of 
churches in the various localities given in the Survey ; while 
on the other hand the actual numbers indicated in some of 
the counties and in special places are evidence enough that 
churches were plentiful. Three hundred and sixty-four are 
mentioned in Suffolk, two hundred and twenty-two in Lin- 
colnshire, one hundred and eighty-six in Kent, one hundred 
and thirty -two in Hampshire. Twenty -four localities in 
Norfolk and sixteen in Kent had more than one place of 
worship apiece ; Norwich city alone possessed fifty-four ; 
Folkestone, Hoo, and Dartford, in Kent, respectively eight, 
six, and four. One manor in Hampshire, that of Chilcombe, 
which is said to embrace eight modern parishes, is credited 
in the Survey with nine churches, a number which would be 
fully up to modern requirements. Postling, in Kent, which 
in its fold of the downs looks as if it had not changed since 
long before Domesday, had at that time two small places of 

One caution must be borne in mind in dealing with 
Domesday evidence. In cases where a church at a certain 
place is mentioned in the Survey and an edifice of early 
character is now to be seen on the spot, the tendency has 
sometimes been to leap to the conclusion that we have a 
Saxon building before us, though there may be nothing about 
it of pre-Conquest character. A period of some twenty years 
elapsed between the Conquest and the taking of the Survey, 
the date of which is subsequent to 1085-6, and Norman 
churches may have been built in the interval. This may have 
been the case for example at Albury and Abinger, in Surrey, 
where Domesday mentions churches, and we find buildings 
of Early Norman date now upon the sites. It is interest- 
ing to know what is the statistical relation between existing 
Anglo-Saxon churches and churches mentioned in Domesday ; 
to know, that is, how many churches that must have been 
standing when the Survey was taken have found a place in 


it. Taking the places enumerated on the map, Fig. 175, 
and leaving out of account the towns, the churches in which 
are difficult to identify by name, we obtain easily more than 
a hundred places that can be identified in the pages of the 
Survey but it is only at forty-five of these places that there 
is indication in Domesday of a church or even of a priest. 
In other words, it appears that fewer than half of the 
existing structures of pre-Conquest date are mentioned in 
Domesday, and some of the most conspicuous Saxon monu- 
ments, such as those in Northants, are ignored in the Survey, 
All the places in this county where Saxon churches now exist 
are mentioned, but the entry ' ibi ecclesia ' is never added, 
though the church at Pattishall, a pre-Conquest example, is 
incidentally referred to in connection with the location of a 
plot of land. 

The presence on a site of carved tombstones and crosses of 
pre-Conquest type may be held to prove that there existed 
there in Saxon times a graveyard and in all probability a 
church. Such monuments however no more prove the 
Saxon date of an edifice in or near which they may now be 
found than does the mention of an ecclesia at some special 
village in Domesday involve the antiquity of its present parish 
church. They do not tell us whether the church by which 
they were originally located was of stone or of wood, though it 
may be noted that their number, which in some parts, counting 
fragments, is very great, is at any rate evidence of considerable 
activity as well as skill on the part of the Saxon worker in 
stone. The ornamental forms and the figure sculpture on 
these stones are of importance in connection with the decorative 
details of the churches, but the comparisons thus suggested 
must be reserved for treatment on another occasion. 

Saxon fonts tell us no more than Saxon tombstones, but 
Saxon sundials have this further value, that, being of stone and 
forming integral parts of the fabric, they imply a church of 
this material. A full list of existing Saxon sundials is a 


desideratum, and would be a document with some significance 
for the stone architecture of the pre-Conquest period. 

If documentary notices be as we have seen fortuitous, in 
that no vahd conclusions can be drawn from them as to the 
total number of Saxon churches, the same may be said almost 
as confidently about the cases of actual survival. No general 
principle seems to be involved in the complete or partial 
preservation of certain examples and the total disappearance of 
others. We cannot say that such and such a percentage of 
Saxon churches has perished or been preserved. We have 
no ground even for saying that the churches which have 
entirely perished owed their destruction to the fact that they 
were of wood, while the stone ones were, as a rule, preserved. 
Local circumstances doubtless determined the treatment of the 
local shrine in the eleventh as in all succeeding centuries of the 
mediaeval epoch. The timber churches that existed, no doubt 
in considerable numbers, at the Conquest were gradually 
replaced by stone structures, and this process which has been 
going on from the earliest Saxon times ^ is not yet complete, 
for the wooden walls of one example are still standing ; 
while not only Saxon wooden churches but many Saxon 
stone churches were pulled down and rebuilt by the Normans. 
At Lastingham in Yorkshire for example there was a Saxon 
stone church,- but the fabric of the present edifice is Norman. 
This process of rebuilding stone churches in a later style has 
been in progress ever since, for Saxon churches, as at Framing- 
ham Pigot, near Norwich, have been replaced quite in our own 
time by modern structures. The cases of complete or partial 
survival are therefore of an accidental or casual kind, and give 
us no help in estimating the actual former wealth in churches 
of Saxon England. 

In connection with this subject, it may be noticed that 
the churches mentioned in the Survey and the charters 
are referred to under the three terms ' ecclesia,' ' ecclesiola,' 
1 Vol. I, p. 167. 2 Jbid, p. 210. 


* capella,' which remind us of the division of churches in the 
law of ^thelred, noticed under the headings ' head minster,' 
' minster of middle size,' ' lesser church with a graveyard,' 
and 'field church,'^ A study of the monuments gives us no 
more help towards fixing any architectural significance for these 
terms than does the study of the records.'- There are in Eng- 
land two surviving examples of the pre-Conquest capella or 
ecclesioia existing side by side with Saxon parish churches, an 
arrangement frequently indicated in the documents.^ One is at 
Heysham in Lancashire where the capella is of primitive form;* 
the other is at Deerhurst in Gloucestershire and here the capella 
is a complete church with a nave and chancel.^ The Saxon 
church at Bradford-on-Avon, Wilts, is referred to by William 
of Malmesbury as an ecclesioia, but though small in size it is 
somewhat elaborate and possessed nave, chancel, and lateral 
porches or quasi-transepts, 

(2) The distribution of the examples as shown on the 
map of Saxon churches suggests some comment. They are in 
the first place confined to England. In the Lowlands of 
Scotland, though some districts of these received at an early 
date an Anglian population, the researches of Messrs. 
M 'Gibbon and Ross^ have not revealed a single example 
with the special Saxon characteristics. It is true that there 
exist buildings in Scotland which seem to have dropped the 
distinctively Celtic motives and yet are not in their features 
characteristically Norman. The tower of Restennet Priory, 
Forfarshire, and St. Regulus' Chapel, at St. Andrews, with the 
upper stage of the round tower at Abernethey, Perthshire, may 
be mentioned in this connection. They may exhibit a pre- 
Norman Romanesque contemporary with Late Saxon work, but 

^ Liebermann, Gesetzc, i, 264. ^ Vol. i, p. 309, note. ^ ibid, p. 3 10 f. 

*The plan of the early chapel at Heysham, with its surroundings, was given 
in vol. I, facing p. 312. 
^postea, p. 1 10. 
* The Ecclesiastical Architecture of Scotland, Edinburgh, 1896-97. 


they show none of the unmistakeable peculiarities which mark 
the Saxon style in England, In Wales M. H. Bloxam claimed 
a pre-Conquest origin for the tower on Priestholm or Puffin 
Island, but this is clearly Norman, of no earlier date, though 
of simpler workmanship, than the tower of Penmon priory 
church on the neighbouring coast of Anglesea. Nothing 
Saxon seems to have been noted elsewhere in the Principality 
or in Cornwall. The church of Tintagel in Cornwall has 
certain features of Saxon character but these are not pro- 
nounced enough for it to be placed in our list of examples. 
The early ecclesiastical buildings in all these parts of 
Great Britain belong to the types generally termed Celtic, 
of which a notice has already been given (ante, chapter I). 
There is one building in Ireland which has been claimed 
as akin to our own Saxon structures, and this is the western 
part of the priory church at Howth, on Dublin Bay. This 
structure however, while it lacks the usual Irish characteristics, 
does not exhibit any of the special features of Saxon buildings. 

Saxon architecture proper is not only confined to England, 
but, as shown on the map. Fig. 175, it is more especially repre- 
sented in the eastern and midland counties. Examples, if 
they exist at all, are very infrequent on the western side of the 
Pennine chain from Cumberland to the Mersey, in Stafford and 
Cheshire, and more to the south in the counties of Monmouth, 
Somerset, Dorset, and Devon. This may, of course, be ex- 
plained in great part by the late and gradual Teutonizing of the 
western parts of the country ; but it is not a little remarkable 
to find in Shropshire a kind of wedge of Saxon architecture 
driven, so to say, into the midst of the district in whose earl) 
ecclesiology Celtic traditions were predominant. The Saxon 
examples in this county invite the conjecture that a systematic 
search in the west of England generally might bring to light a 
good many more. The south-western counties probably con- 
tain more examples than have as yet been noticed. 

The suggestion just made has wider bearings. Taking not 


England as a whole, but the smaller areas of the eight different 
Districts into which for convenience sake the map has been 
divided, we note a tendency among the examples to fall into 
groups, while pretty wide regions are on the other hand left 
blank. The explanation partly is that when one example in a 
certain district is brought to light and commented on, the 
interest thereby excited leads to the recognition of other 
examples of a similar style of work in the neighbourhood. 
A group is in this way formed in one district, while in another 
the initial discovery still remains to be made, and the ground 
is in the meantime barren. It would be a mistake therefore 
to attach too much significance to the actual distribution of the 
monuments on the basis of our present knowledge. These 
geographical statistics must be regarded as to some extent pro- 
visional, and the barren regions should in the meantime be 
regarded as places where local investigation is specially called 
for. The question of the geographical distribution, not of the 
churches themselves simply as churches, but of the special types 
of churches and of their distinctive constructional or artistic 
features will be noticed later on. 

(3) With regard to the criteria relied on for the separation 
of Saxon churches from those of other architectural periods, 
the question in most cases resolves itself into one between Saxon 
and Norman, and accordingly characteristics which are found 
in both Saxon and Norman buildings are not of much help in 
the discrimination of doubtful examples. Features which do 
not occur in Norman work on the Continent are the most 
valuable for the purpose in view, and the appearance of them 
is enough either to stamp a structure as pre-Conquest, or 
else to prove the survival or occasional recrudescence of native 
English traditions among Anglo-Norman builders. The word 
' pre-Conquest ' must of course be taken to refer to style rather 
than to actual date. Just as the earliest work at Westminster 
Abbey is Norman though executed before 1066, so buildings 
that are still essentially Saxon may have been actually reared 



after the accession of William. On the other hand some 
specially Saxon features do make their appearance sporadically 
in work that must have been done, if not by Norman hands, 
yet at Norman bidding and on Norman design. There is for 
example a distinct ' long-and-short ' feeling about the massive 
quoins of the fine Norman gateway at Rougemont, Exeter, 
and in the jambs of a Norman doorway of the nave at Stow, 
Lincolnshire ; the double-splayed window, a distinctly non- 
Norman feature, occurs in what must be Norman work on the 
west side of the cloisters at Norwich ; some chancel arches 
that are Norman in technique have their jambs of a Saxon 
plan. Such instances occur and must not be ignored. They 
are however not at all numerous and it would be a mistake to 
make them of much importance. They do not invalidate the 
general distinctions of style already laid down. 

Any detailed discussion of the special features on which so 
many of these criteria depend would be useless unless the reader 
were possessed of some general knowledge of the buildings in 
question. It will therefore be on the whole the most con- 
venient procedure to Introduce the subject as If those were 
addressed to whom Saxon architecture is as yet only a name. 
When the main facts have been made clear by description and 
illustration, the time will have come for an analysis of them 
from the points of view of origin, date and continental affinity. 

Are there any general criteria by which an Intelligent observer 
can distinguish a Saxon church from one belonging to the other 
mediaeval periods ? There is no criterion of absolute validity 
but there are general symptoms which can be diagnosed even 
from the bicycle saddle. 

The first sight of a country church is generally of its tower 
and spire, A western tower that is of great height In propor- 
tion to its width and of conspicuous plainness will repay 
interrogation. If it be buttressed at the angles it is no use 
inquiring further, unless indeed the buttresses can be plainly 


seen to be later additions. If it rise gaunt and smooth, the 
outline only broken perhaps by a single horizontal string- 
course above which it may slightly narrow, it has Saxon 
character and it is worth while devoting special attention to 
the belfry openings. If these be recessed the tower is almost 
certainly of Norman or later date, but if they possess the 
special characteristics already indicated in connection with 
German work (ante, p. 62 f.), the building may be set down 
as in all probability of pre-Conquest origin. 

Only less conspicuous than the tower of a church is the 
general shape of the body of it. When the side walls of the 
nave are of great proportionate height there is a suggestion of 
early date, but the appearance is often deceptive. There are 
Early Norman churches of remarkable height of walls, and 
furthermore the height is sometimes merely due to the addition 
in late mediaeval times of a Perpendicular clearstory on the top 
of an earlier wall. On the other hand, when further investiga- 
tion shows that the nave is of very great proportionate length 
as well as height, a strong presumption of Saxon origin is at 
once involved. It would be a mistake however to imamne 
that all Saxon naves are long, or that there is any one scheme 
of proportion that is exclusively the possession of the style. 
The analysis of the proportions of a number of Saxon naves 
the result of which is given in Fig. 30 proves the contrary. 

It will be seen that some examples, notably in Kent, are 
com.paratively wide, while others, especially some in the north 
like Monkwearmouth and Escomb, are long and narrow, 
while various intermediate degrees of proportion are also 
represented. The evidence of the extraordinary length of 
Jarrow old church is the statement in Hutchinson's History 
of Durham^ 11, 475 f., that it measured twenty-eight paces by a 
width of only six. The walls were nearly thirty feet high. 

The presumption of Saxon origin based on elongated 
proportions is strengthened if there be any indications 
that the original pitch of the roof was a steep one. If 







R OC H ESTER ^^^^^ 








,— bRf 

,71^" is] 

53. b* 16 

ctSK 17.6 

U K 29* 

K-O.t X IS 

J7.6X Mi 
36.6 >i.l7 

3V x H 

iS. * 17 


2.7 « If. 

^..6> 16.6 

I « 13 

Fig. 30. — Comparative chart showing the proportions of the naves (interior 
measurements) of twenty-four Saxon churches. 










the original gable be not preserved the mark of it is some- 
times seen on the wall of the tower, and a sharp-pointed gable 
is a Saxon peculiarity. If the character of the masonry be then 
examined some confirmation of the hypothesis of a Saxon origin 
may be found in the comparative rudeness and irregularity of 
the technique and the absence of any special treatment of the 
face, such as herring-bone work. Herring-bone work, which 
used to be considered a 
sign of Saxon origin, is now 
known to raise a presump- 
tion to the contrary. More 
assurance will be gained if 
the thickness of the wall 
turn out to be comparatively 
slight, say from 2 ft. to 2 ft. 
6 in. Comparative thinness 
of wall is a good but by no 
means an absolute test of 
Saxon and Norman, and this 
measurement should always 
be taken. Norman walls 
nearly always run thicker 
than Saxon. 

It may be asked whether 
the presence of all these indications in proportion, technique, 
thinness of walling, etc., be not enough to prove pre-Conquest 
origin ? The question can hardly be answered in an absolute 
form. Every investigator must rely to a certain extent on 
his personal judgment the exact grounds of which cannot 
always be strictly formulated. The general aspect of a 
structure counts for something in any decision as to its date 
and style, and this can only tell on actual inspection. The 
local position and surroundings of a building must also be 
taken into account, and certain kinds of evidence are of 
more weight in one part of the country than in another. 


31. — Plinths of Anglo-Saxon walls. 
(Scale, c. ^ in. to i ft.) 

A. Dunham Magna, Norfolk. 

B. Hainton Church tower, Lincolnshire. 

C. St. ^Martin, Wareham, Dorset. 

D. Clee, Lincolnshire, plinth to tower arch. 

E. Stow, Lincolnshire, south transept. 


There is no question however that definite features and bits 
of characteristic detail are of greater value as criteria than the 
general aspect of a structure, and in the list of monuments 

V /f. ■= r^- 


Fig. 32. — Quoin of nave, St. Mildred, Canterbury. 

which forms the basis of the map, Fig. 29, reliance has 
been placed almost exclusively on these definite features, and 
not on the more general considerations. 

Taking these features and details therefore in order, we may 
note first that a Saxon wall may or may not have a plinth or 
base-moulding. There are different forms of these plinths in 



pre-Conquest work but no one of them is exclusively Saxon. 
Fig. 3 1 shows some examples. It will be noted that the profiles 
of the members are either square or chamfered, and this rustic 
simplicity in forms is very characteristic of the Saxon style. 
A moulded plinth with the 
classical cyma is to be seen at 
the base of the jambs of the 
tower-arch at Alkborough, 
Lincolnshire, but the work is 
Roman re-used. 

The most important point 
about the wall is the treatment 
of its corners. There are two 
characteristic Saxon methods 
of quoining, by the use of big 
stones and by what is termed 
long-and-short work. Great 
squared stones measuring a 
yard or more in their greatest 
dimension, massed one on the 
other and lending to the 
corner a rock-like solidity, are a feature found in some Saxon 
churches but not in later mediaeval work. Fig. 32 shows 
these big stones in the south-west quoin of the church of 
St. Mildred, Canterbury. The size of the lowest stone is 4 ft. 
in height by 2 ft. 8 in. in width and i ft. 5 in. in thickness. 
A more regular use of such stones is seen in Fig. 23 from 
the south-west corner of the southern transept at Stow, 
Lincolnshire. They are shaped somewhat like huge bricks 
and laid with their longest dimension along the southern and 
western walls in alternation. They run about 2 ft. to i ft. 
6 in. in height. 

Long-and-short work is the best known and most easily 
diagnosed of all Saxon symptoms. The following is the method 
of it. An upright pillar of stone, square in section and in 

Fig. 33. — Quoin stones, south-west 
angle of south transept, Stow, 



height varying from about 2 ft. to 4 ft., is placed at the angle 
of the structure, it may be that of the tower, the nave, the 
chancel, or the porch, and over it is laid a flat slab of stone 
which grips into the wall and shows the length of its sides 
along the two faces. The correct designation of the work 
when the whole of it is seen would be ' upright-and-flat ' 

rather than long-and- 



short, but the latter 
term becomes appro- 
priate when, as is often 
the case, the surface of 
the walls is plastered. 
With a view to plaster- 
ing, which was a common 
perhaps a normal finish 
to Saxon walling, the 
wall-face was set back 
some half inch or so 
from the surface of the 
upright stones on the 
quoins and the plaster 
brought up flush with 
the edge of these. The 
parts of the flat slabs 

Fig. 'iA.. — South-east quoin of nave at Witter- 111 1 n 

ij , 11 J 1 . that lay along the walls 

ing, JNortnants, showing long-and-short work. ^ '-' 

were cut back level with 
the wall-faces and covered with the plaster so that only that 
portion of them was visible which corresponded with the 
width of the uprights. This portion was in height only the 
thickness of the slab and appeared ' short ' in comparison with 
the ' long ' upright pillars. Fig. 34 exhibits the technique 
in an example where the plaster has been stripped from the 
stonework. Were the plaster present the tailing of the flat 
piece into the wall would not be seen. This quoining is an 
excellent Saxon criterion for, so far as the writer's knowledge 



goes, it is never used in Norman work in the Duchy or 
indeed anywhere on the Continent/ though there are 
occasional survivals of it in Norman work in England. 

An equally significant feature and one that can be caught 
in a passing glance is the so-called pilaster strip. This must 
be carefully distinguished from the buttress. Saxon walls in 
general are unbuttressed, though the buttress occurs in three 
Kentish examples, St. Martin and St. Pancras, Canterbury, 
and Reculver ; also in the enigmatical church of St. Peter- 
on-the-Wall, Essex, and round the apse at Brixworth. These 
buildings are all presumably early. 

Fig. 35. — Comparison of Saxon pilaster strips at Earls Barton tower (A), with 
Early Norman corner buttresses at West Mailing Church, Kent (B). 

The buttress in various forms is common in Norman and 
later architecture, but its place in Saxon work is filled by the 
pilaster strip. The pilaster strip accords with the German 
Lisene in that it is not meant like the buttress to add 
strength to the wall. It is too narrow and too slight in 
projection for any purpose of the kind, but is simply a 
flat upright band of stonework varying in width from 4 in. 
to I ft. and running at intervals up the wall for the sake 
of decorative efl^ect. The fact that the Saxon pilaster strip is 
sometimes found ascending a wall above the crown of an arched 

^ That is to say for quoins. There are instances in Germany, as, e.g. 
on the west front of St. Pantaleon at Cologne, where long-and-short 
technique is used for wall-pilasters, i.e. in the manner shown in Fig. 37. 


doorway (see postea, p. 182) is conclusive proof that it has no 
constructive significance. The difference between the decorative 
pilaster strip and the strengthening buttress is seen by compar- 
ing in Fig. 3 5 the plan of the 
face of a Saxon tower A with 
that of the face of an Early 
Norman tower B. The flat, 
broad, clasping corner but- 
tresses of the latter are as 
characteristically Norman as 
the narrow strips only about 
4 in. in width of the former are 
characteristically Saxon. The 
sort of appearance presented 
by such strips can be seen in 
Fig. 36, which shows the south 
side of the nave at Woolbeding, 
near Midhurst, Sussex. The 
pilaster strips are irregularly 
spaced and are 7 in. wide. Such 
strips areHat times, but not 
always, constructed in the long- 
and-short technique as shown 
in Fig. 37. The appearance 
of these features which should 
not be more than 10 in. or i ft. 
in width is the most certain 
test we possess that the piece 
of walling where they appear is 
of Saxon origin. 
A test on which much reliance may be placed is that of windows. 
A distinction may be drawn here between the wide openings 
common in belfries, and the narrower apertures for light in walls. 
The Saxon belfry opening (see ante, p. 60 f.), which also 
occurs occasionally in the nave walls of pre-Norman churches, 


37. — Pilaster strip, Breamore, 
Hants. The strips are about 
1 1 in. wide. 


Fig. 36.— Pilaster strips on the nave wall of Woolbeding Church, Sussex. 
The strips are about 7 in. in width. 

(To face p. 90.) 



is cut straight through the whole thickness of the wall, 

and it is subdivided generally into two.^ The way in 

which the partition is managed is the point to note. Each 

half of the aperture is covered by a small round arch and 

between these a bit of the wall would be left suspended in the 

air were it not held beneath 

by a flat stone slab of sufficient 

area, that is itself sustained by 

a single prop in the form of a 

stone shaft that stands under 

the centre of it. The form 

of the shaft and of the cap 

which sometimes surmounts 

it are often worthy of remark. 

The former is sometimes that 

of a plain cylindrical column, 

sometimes it is octagonal, or 

square with rounded corners. 

It takes however occasionally 

the peculiar form of a baluster 

with a series of projections Fig. 38. — Subdivided opening with 

and hollows that appear as if baluster shaft in the tower of 

formed on the lathe. The Barton-on-Humber, Lincolnshire. 

. , 1 • 1 The splaying downwards of the 

caps are either cubical or ... . , 1 r ■ 

'^ sill IS a later modmcation. 

ionic of a debased form. 

Both shafts and caps will be studied on a subsequent page. 
It is the combination of the unrecessed openings, the flat slab 
called a through-stone, and the shaft called from its position 
a mid-wall shaft, that is the distinguishing mark which gives 
the whole structure a Saxon stamp. Such windows in the upper 
story of a tower and in the side wall of a nave are shown in 
Figs. 38 and 39. 

^ There is one instance of a triple opening of the kind, though not in a 
belfry, at Brix worth, Northants, and the belfry openings in Earls Barton 
tower in the same county are fivefold. All others are double. 


In regard to the narrower single apertures for light, a 
window with a very narrow external opening the rounded 
head of which is cut out of a single stone, and a very wide 
internal splay, is often pronounced Saxon whereas there is an 

f ic «L of l*/4ll. 

--20" > 

-2;"-;- ■. 

f»ce of str'tna cowje 

Fig. 39.— Window in north wall of nave, Worth, Sussex, with plan. 

equal or greater likelihood that it is Norman. There are it is 
true Saxon windows of this general form, both early and late 
in the style (postea, p. 273), but the really characteristic Saxon 
light is one in which the narrow aperture is in the middle of 
the thickness of the wall and there is a splay on both sides 
(Fig. 40). These double-splayed windows may be either 
circular or of the upright round-headed shape, and the actual 



opening for light is at times cut in a thin slab of stone 
or plank of wood built into 
the wall at the centre of its 
thickness. The plans of a 
few characteristic Saxon window 
openings with single and double 
splay are given in Fig. 41. Too 
much should not be made of flat 
heads to windows or doors (see 
postea, Figs. 6^ and 66, p. r 14), 
or heads made of stone slabs 
set up against each other at an 
angle, and forming what is some- 
times called a straight-sided arch 
(Fig. 42). Heads of openings 
of the kind occur in Roman ^ 
and in Norman ^ work as well as 
in Saxon. On the other hand a peculiarity that suggests a 
Saxon date is the inclination of the jambs of door or window 

Fic;. 40. — Double-splayed Saxon 
window at DIddlebury Church, 
Shropshire. From a drawing by 
the late J. T. Irvine. 

Fig. 41. — Comparative plans of window openings. 

7. Brixworth, east wall. 

8. Boarhunt, Hants. 

9. Barton-on-Humber, western adjunct, 
10. Tower light, Howe, Norfolk. 
II and 12. Ledl)ury, Herefordshire, and 

Overbury, Worcestershire (not drawn 
to scale), of Norman date. 

1. Clearstory at Brixworth. 

2. St. Martin, Wareham. 

3. Splayed Roman at Cilurnum on the 

North Tyne. 

4. West Wall of Church, Monkwearmouth. 

5. Killiney Church, Ireland. 

6. Chancel at West Hampnelt, Sussex. 

openings by which the aperture is made narrower above than 
below. This characteristic, which we have seen to occur in 

^ Baths of Car.ica!!,!. 

-Jarrow, Durham. 


Ireland, generally points to a pre-Conquest origin. An example 
from Brigstock church, Northamptonshire, is shown in Fig. 43. 
The stranger who is interested in our early buildings, and is 
attracted by these external indications to enter the church, 
should notice the position and character of the doorways. It 
is characteristically Saxon to place these north and south, often 

Fig. 1^42. — Opening in tower at 
Brigstock, Northants. From 
a drawing by the late J. T. 

Fig. 43. — Window with sloping 
jambs, at Brigstock. Do. 

just opposite each other, at the western end of the nave. One 
of them, generally the north, is now very commonly blocked. 
Narrow proportions in relation to height are Saxon. Fig. 44 
shows a characteristic specimen. 

In the interior of the edifice, if it be an aisled church, indica- 
tions may often be found that the existing arcades have been 
cut through an earlier wall. The original windows of the aisle- 
less nave may be partly preserved and it is worth while peering 
for the marks of their once external openings, on the aisle side 
of the wall of the nave between the arcade and the roof of the 
aisle. If the arcade be a Norman one the conclusion is some- 
times drawn that the earlier wall must be Saxon. A good deal 
depends, however, upon the date of the Norman work, for it 
is quite possible that an Early Norman wall may have been cut 



through in later Norman times, just as, to compare great things 
with small, the choir of Lan- 
franc at Canterbury was re- 
placed within the Norman 
period by the grander structure 
of Ernulf. This question was 
posed by the late Professor 
Freeman in a paper on Iver 
church, Bucks,^ where it was 
found that the wall above a 
Norman arcade of the first 
part of the twelfth century 
was of earlier date, and he 
regarded it as inherently 
possible that in the same wall 
there might be two periods 
of Norman work separated 
by no long interval of time. 

=Si^— J!tJ.hs/Ai3. 

Fig. 44. — Saxon doorway at Diddle- 
bury Church, Shropshire, From a 
sketch by the late J. T. Irvine. 

Iver possesses the remains of some of the original windows 
of the once aisleless nave, now blocked and cut into by the later 

arcade (Fig. 45). The treatment 
of the window heads has been 
held to be proof of their Saxon 
origin, and they were so accepted 
by Professor Freeman. The 
manner in which the mouldings 
are stopped on the jamb, which 
itself is perfectly plain save for 
the roll on its edge, can however 
be paralleled so nearly in an 
Early Norman doorway at St. 
Nicholas, Caen, that Iver church 
has been excluded from the list of Saxon examples. It is 
unfortunately impossible to secure the plan of the jambs 

^ Archaeological Journal, vii. 

Fig. 45. — Blocked window, Early 
Norman or Saxon, in the 
nave of Iver Church, Bucks. 


which might have settled the question. The window was 
probably not double-splayed, but one of the type of Nos. 1 1 
and 12 in Fig. 41. 

Fig. 46. — Tower arch, Market- Fig. 47. — Chancelarch at Stainton- 
Overton, Rutland. Saxon by-Tickhill, Yorks. Norman 

technique. technique. 

The principal features in which Saxon character may be 
expected to show itself in an interior are the tower and 
chancel arches. The first point to observe here is the masonry 
of the jambs and arch. If the stones of which these are 
composed run through the whole thickness of the wall there is 
every probability that the work is pre-Conquest, but if there be 
squared stones on the two faces of the wall with rubble filling 
In the middle, the probability is very greatly decreased. This 



difference, which is one of no little significance, is illustrated 
by the arches placed together for contrast in Figs. 46 and 47. 
An arch built with through-stones as at Market Overton is 
certainly either Saxon or Roman ; one of the other kind, though 
more probably Norman, may yet be of pre-Conquest date. 

Fig. 48. — Tower arch, Brigstock, Northants, showing pilaster strips carried 
round the arch. The pilaster strips on each side start from projecting 
corbels. From a drawing by the late J. T Irvine. 

In this case observation should be made of the method of 
cutting the voussoirs of the archivolt, of the form of the im- 
posts, and of the enrichment, if any exist, on one or both of the 
faces of the arch. The signs of ignorance or want of skill in 
shaping and fitting the wedge-shaped voussoirs, and a clumsy 
or a fantastic character in the imposts, are Saxon symptoms ; but 
the most significant feature, where it occurs, is the use for the 


enrichment of the arch of pilaster strips like those on external 

A B 

Fig. 49. — Jambs of western doorways. 
A. Kirk Hammerton, Yorkshire. B, Kirkdale, Yorkshire. 

walls. These flank the opening and are bent round above in 
the position of a hood moulding after the manner illustrated in 

Fig. 48. The projecting corbels 
from which they start at the 
bottom are noteworthy. This 
strip-work round openings is a 
very good criterion of Saxon 
origin. A recessed arch, and one 
with angle shafts, and with the 
edges of the openings worked 
into roll mouldings may be late 
Saxon but are more likely to be 
the work of Norman hands. A 
pair of Saxon recessed arches 
with the angle shaft is shown in 

Fig. 50. — Portion of south door to 
nave at Stopham Church, Sussex. 

Fig. 49 and the use of the roll moulding is illustrated in Fig. 50. 



From these preliminary questions about the number, dis- 
tribution, and criteria of Saxon churches, we may now pass 
to an analysis of the various types which they present from 
the simpler to the more complex. What is aimed at here is 
chiefly description. The object in view is to put the reader in 
possession of the main facts concerning the plan, the general 
form, and the details of the most characteristic examples of 
Saxon ecclesiastical architecture. Discussions of continental 
affinities and comparative dates are for the most part reserved 
for a subsequent chapter. It is proposed to perambulate the 
Districts shown on the map and to describe whatever is 
of special interest in each,^ but, in order to present as full a 
view as possible of the subject, each type in succession will be 
followed over the country wherever it presents itself in a 
characteristic example. This will give an idea whether the 
type is a rare or a common one and will exhibit the geo- 
graphical limits of its distribution. 

Bv following the subjoined order we shall have the oppor- 
tunity of passing in review the principal types of Saxon 
churches as well as the larger parts and features which 
distinguish them. 

^ Only a limited number of the examples on the map are discussed in 
the text. A full list, with a brief indication of what appears in each, will 
be found at the end of the volume. 


There will be noticed then 

I. The plain rectangular oratory without chancel. 
II. The same with rectangular chancel. 

III. The same with apsidal presbytery. 

IV. The oratory with space screened off before the apse. 
V. The western or lateral porch. 

VI. The lateral chapel. 
VII. The western tower. 
VIII. The tower forming the body of a church. 

IX. Axial towers. 
X. Central towers, transepts, and the cruciform plan. 

XI. Churches with both a central and a western tower. 
XII. The twin-towered facade. 

XIII. The triple-apsed plan. 

XIV. Aisled churches, 
XV. Crypts. 

I. The Plain Rectangular Oratorv without Chancel. 

The simplest type of church that England has to show 
is the oblong interior without divisions, a form familiar as we 
have seen to Irish traditions and represented by existing 
remains in all the Celtic districts of the British Isles. Apart 
from, some instances in Cornwall that are certainly not Saxon,^ 
the writer knows only one example in the English counties that 
is of assured pre-Conquest date, and there is some doubt even 
about this whether it is Saxon and not rather a specimen of 
Irish handiwork. The building in question is the so-called 
chapel of St. Patrick upon a promontory of the Lancashire coast 
at Heysham on Morecambe Bay. It stands in a ruined con- 
dition on a rocky table a few yards west of the parish church of 

1 Of Cornish examples the best known is the oratory of St. Piran on the sand 
dunes a couple of miles north east from Perranporth. It measures internally 
25 ft. by 12 ft. The principal door was on the south and there was a smaller 
one in the east wall. 



Heysham which is itself in part of Saxon date, and has in front 
ot it a small sanctuary used in mediaeval times as a place of 
burial. Could we be sure that the marks of interment go back 
to pre-Conquest times it could be claimed as an example of the 
ecclesiola or capella cum coemiterio referred to in the previous 
volume.^ The plan of the chapel is shown in Fig. 5 i in con- 

FiG. 51. — Plans of single-celled oratories. 

A. Oratory on Skellig Michael, rounded within but square outside. 

B. Rectangular Oratory on Skellig Michael. See Fig. 9 ante, p. 22. 

C. Foundations of rectangular Oratory near Aber, North Wales. See T/ie Builder, 

Oct. 2, 1897, p. 252. 

D. Ileysham Chapel, Lancashire. 

nection with one or two plans of characteristic Celtic single- 
celled oratories of the same type, with which it may be 
compared. It exhibits great length in proportion to the width, 

^ Vol. 1, p. 311, and Fig. 25. A pre-Conquest date has been claimed for 
the graves hollowed in the rock to the west of the chapel, on the supposition 
that some marks on the flat stone to the north of the northernmost of the 
cavities are the remains of an ornamental pattern of interlaced work similar to 
that on pre-Conquest stone crosses. As a fact these are only pick marks made 
when the surface of the stone was dressed flat. Marks exactly similar can be 
seen on the walls of the square sinkings where the head-stones stood. 



the internal dimensions being about 27 ft. in length by a width 
that varies from nearly 9 ft. to less than 8 ft. The wall is 
about 2 ft. 5 in. thick. There is no trace of an altar and no 


Fig. 52. — South doorway of the Chapel Fig. 53. — Doorway in north wall of nave, 
of St. Patrick, Heysham, Lancashire. Somerford Keynes, Wilts. 

sign of a window in the east gable which is well preserved, 
but the marks of one are visible in the south wall. The 
only features to be noted are, first, a curious projecting stone 
at the base of the east gable on the north side, that may be 
compared with the Irish stone of similar form shown in Fig. 
17, ante, p. 31, and the south doorway, the scheme of which 


is given in Fig. 52. It will be seen how the stones are 
arranged in the jambs. Upright slabs which go through the 
whole thickness of the wall alternate with smaller slabs laid 
horizontally. The head of the doorway is cut out of a single 
stone and is enriched with hollow flutings. There is a some- 
what close parallel to this ruined doorway in a Saxon one 
better preserved, but blocked, in the north wall of the church 
of Somerford Keynes in Wiltshire, shown in Fig. 53. The 
mouldings in the latter case project and are ornamented with 
a cable pattern, but otherwise the two correspond so nearly 
that we are inclined to claim for the building at Heysham 
a Saxon rather than an Irish orip-in. 

There is a peculiarity about this doorway at Heysham 
chapel that can be discerned on the plan, D, Fig. 51. The 
jambs have a shallow rebate on their internal face for a door, 
and this in Saxon work is extremely rare. As a general rule 
the doorways are cut straight through the wall and the door 
shut flat across the inner aperture. The doorway at Diddle- 
bury shown in Fig. 44, ante, p. 95, still preserves on the inner 
face the iron hooks on which the door was suspended across 
the opening. The doorways of early date at Brixworth, 
St. Pancras, Canterbury, and of a late epoch at Barnack, Earls 
Barton and Barton-on-Humber, with very many others, are 
planned in this fashion, but it is not the case that all Saxon 
door openings are thus treated. At Monkwearmouth porch 
the north and south doorways are rebated for doors opening 
outwards from the porch (see plan. Fig. 78, postea, p. 141) and 
the feature here is unquestionably original. The north door 
of the nave at Escomb has also a rebate, and so probably had 
the two other doors of the church. (See plans, Figs, 66 and 
62, postea, pp. 114, I 10.) At Reculver, the original and very 
early doorway into the space before the apse has the jambs in 
the interior cut into a shallow rebate 2 in. in depth by a length 
along the wall of 6 in. 



II. The Oratory with Rectangular Chancel. 

The succeeding types exhibit a distinct sanctuary added 
to the original oblong which now becomes the nave. When 
the sanctuary is of rectangular form we have a type which is 
familiar in Ireland and in Scotland and is among those most 
commonly represented in the Saxon monuments of our own 
country. The following are some characteristic examples 
from the different districts. In District I Whitfield, near 
Dover, Kent, presents to us, as the core of an extended and 
modernized building, a small nave and chancel church of the 

Fig, 54. — Plan, and section of nave, of Whitfield Church, near Dover, Kent. 

plan and section shown in Fig. 54. The chancel arch has 
been widened in modern times. We note the minute 
dimensions of the chancel, only 9 ft. from west to east, and 
the great proportionate height of the side walls. These are 
2 ft. thick in the nave but in the chancel only i ft. 9 in. 
In the south wall of the nave is a small original window 
more than 10 ft, above the floor the aperture of which, 8 in. 
across, is in the middle of the thickness of the wall, while 
the jambs are splayed out towards either face, to an opening 
of about 2 ft. 3 in. in total width. These double-splayed 
windows we shall constantly meet with as we proceed in our 

One of the most complete Saxon churches of this type is 
Boarhunt in Hampshire. The plan is given in Fig. §^ and 
a general view of the exterior in Fig. 57 which shows that 



unlike the last example the walls are by no means of abnormal 
height. The chief peculiarity of the plan is the former existence 

55. — Plan of Boarhunt Church, Hants. 

of a cross-wall cutting off a portion of the western end of the 
nave, after a fashion observed also at the Saxon church at 
Daglingworth in Glou- 
cestershire. Of features 
there are to be noted, 
first, the double-splayed 
north window of the 
chancel. Here the aper- 
ture is cut in a mid-wall 
slab set in the centre 
of the thickness of the 
masonry, and as will be 
seen in Fig. ^6 this has 
round it just where it is 
set in the wall a carved 
moulding of the cable 

pattern, that is commonly ^'*^- S^. — Window in chancel at Boarhunt 

used in this country in ^^""''^^ "'"''• ^""^^^ '"^ '^"^'"S ^r 

, , . ^ „ ^ the late 1. T. Irvine. 

work that is of Roman of 

Saxon of Norman and apparently also of comparatively modern ^ 

date. Interesting peculiarities are the marks of staples on 

^ Cable moulding occurs on the imposts of the chancel arch in Heysham 
parish church, where it is most probably seventeenth century work. 

oo^^^T- :■:> -5cirot^ir^*-5i. 


the jamb where hung in old times the shutter which was 
the only means of closing up the aperture. There is a 
rebate round the opening into which this shutter fitted, and 
this becomes a splay above to admit of the easy action of 
the shutter. Another feature of interest is the chancel arch, 
but the character this shows is better represented in other 
examples. The old north and south doors are blocked. 

It will be noticed in the general view of the church that 
a pilaster strip runs up into the central point of the gable 
on its eastern face. This feature is sufficiently peculiar to 
make it probable that all examples of it will be of about 
the same date. Nov/ it occurred in the western srable of 
the church at Kirkdale near Kirby Moorside in Yorkshire,^ 
where is to be seen the lapidary inscription noticed in the 
previous volume- which gives the date of the building at 
about the year 1060. The date of Kirkdale being known, 
we can in this manner arrive at a pretty sure approximate 
date for Boarhunt and for other churches with the same 
peculiarity, as well as (with more chronological latitude) for 
the use of pilaster strips in general. These pilaster strips, 
it may be remarked, are very common in the Saxon churches 
of Hampshire and of Sussex and Surrey as far east as the 
longitude of Brighton, but are not found in Kent. They 
are abundant in the midland districts but less common 
in the north. The illustrations that follow will show a con- 
siderable number of examples in various parts. 

Resuming the consideration of the nave and chancel type, 
and passing on to District II, we find at St. Martin, Ware- 
ham, Dorset, a church also in all essentials complete, though 
the original plan has received additions on the north and 

1 The feature is now destroyed owing to the fact that in 1827 the western 
gable was lowered and a tower built up agamst it. It is however distinctly 
shown in a drawing of 1821 published by C. L. R. Tudor, Kirkdale Church, 
Lond. 1876. 

2 Vol. I, p. 356 and Fig. 27. 














west. The building, which stands on the earlier Saxon ramparts 
of the town, is picturesque in outline, and shows long-and- 
short work at the quoins. In contrast to Boarhunt, Wool- 
bedino;, and the group of churches round Winchester, which 
are of low proportions, St. Martin has the lofty walls which 
on the whole are a pre-Conquest characteristic. 

The Midland District IV contributes the Northampton- 
shire example of Wittering. Only the nave and chancel 
are Saxon, the north aisle and the western tower and spire 
being later additions. The church, of which the plan is 

Fig. 58. — Plan of Wittering Church, Northants. 

given in Fig. 58, stands on high ground quite away from 
the present village and is a conspicuous object on the left 
of the Great North Road as we follow its course towards 
Stamford. Wittering exhibits careful long-and-short work 
at the four quoins of the nave and the two of the chancel, 
see Fig. 34, ante, p. 88, and though no original openings are 
preserved it retains a chancel arch that is one of the best 
specimens of the style. The arch has a height of 14 ft. 
6 in. to the crown with a width between the jambs of 
about 7 ft. Fig. 59 gives a view of the northern jamb 
from the west, together with its plan. The impost of heavy 
trapezoidal form takes the eye at once as something quite 
unlike what the Norman and later mediaeval styles have 
to show. It is composed of several pieces, the front part 
being formed of two superimposed slabs each measuring 4 ft. 



from west to east by a depth of 
of about 9 in. They are carried 

Fig. 59. — View and plan of northern 
jamb of chancel arch, Wittering, 
Northants. Scale of plan ^ in. to I ft. 

2 ft. 6 in. and a thickness 
on a jamb that is set back 
6 in. from the face of the 
impost and the plinth below 
so as to allow space for a 
half-round soffit shaft and 
half-round shafts on the 
west and east faces, to which 
correspond roll mouldings 
round the arch above. 
Further out than these there 
is a square strip like an 
external pilaster strip which 
is bent round above to form 
an outer order. The timid 
suggestion of bases and caps 
is very characteristic of 
Saxon work. The imposts 
of the tower arch at Market 
Overton, Rutland, Fig. 46, 
ante, p. 96, are of similar 
character but the jambs there 
are quite plain though com- 
posed in orthodox Saxon 
fashion of through-stones. 

There exists a fair number 
of examples of imposts of 
chancel arches ornamentally 
treated. Some specimens 
of these will be noticed on 
a later page in connection 
with carved capitals. 

It is worth while to 
pause a moment at this 
chancel arch of Wittering, 














for though in some respects it is exceptional, in others it 
represents a type, specimens of which recur constantly in 
pre-Conquest work. It is a Saxon scheme to set three half 
columns on the jamb, one as a soffit shaft and the others on 
the western and eastern faces, these last two being sometimes, 
as at Bosham, Sussex, recessed to receive the columns as angle- 
shafts. The simpler scheme, without recessing or only with 
what looks like the beginning of recessing, is well represented 
at Clayton, near Brighton (Fig. 60). There 
are here no bases or caps, which at Wittering 
appear in embryo and in other examples are 
fully developed. As a sort of outer order 
a square pilaster strip, as at Wittering, is not 
uncommon, and this occurs in many examples Fig. 60. — Jamb 
such as the famous tower arch at Stow, of chancel arch, 

Lincolnshire, the tower arch at Skipwith, „ '^^ °"' "^^^^' 

. ^ Scale ^ in. to I ft. 

Yorks ; Barnack ; St. Benet, Cambridge, etc. 
Rolls and square mouldings round the arch generally corre- 
spond to the half-columns and strips below. It is a Saxon 
peculiarity however that these latter do not always spring from 
a solid plinth as at Wittering, but from projecting corbels, by 
which they are held as it were suspended a few feet above the 
ground. Stow is a marked example, but the peculiarity is 
well seen also in the archway at Brigstock shown in Fig. 48, 
ante, p. 97. 

In District V the Gloucestershire example of Coin Rogers, 
near the Fosse Way a few miles north-east from Cirencester, 
is also complete. Deerhurst Chapel, though it has lost its 
original eastern termination, is too interesting an example 
to be passed over. Deerhurst, near the Severn, eight miles 
north of Gloucester, has been referred to as one of the two 
places in England that still possess a pre-Conquest chapel 
in the same village as a partly Saxon parish church. In 
the year 1675 there was discovered at Deerhurst an inscribed 
stone, now preserved at Oxford, that records the dedication 



by Earl Odda of what he calls a ' royal hall,' by which we 
may understand a ' basilica ' or church, at a date in Edward 
the Confessor's reign that corresponds with April I2, 1056.^ 
This was formerly supposed to refer to the priory church of 
St. Mary, now the parish church, itself a Saxon building of 
exceptional value, but in 1885 the discovery was made of 
a small Saxon chapel incorporated in the fabric of an old 
mansion, now a farm-house, and it is recognized that the 
inscription, which was actually found close to this house, refers 
to the chapel, the date and character of which are accordingly 
fixed. The building is still attached at its eastern end to the 
farm-house as shown in Fig. 61, It has double-splayed 
lights, the late date of which feature is thereby indicated. 

Fig. 62. — Plan of Escomb Church, Durham. 

For the North, the remarkable example of Escomb, Durham, 
in District VIII will suffice to show the even diffusion of this 
type over the country at large. Escomb is by far the most 
interesting of all the specimens of this form of church, and is 
one of the half-dozen or so of extant buildings to which an 
early date in the Saxon period can confidently be assigned. 
The plan. Fig. 62, shows a nave of remarkable length in 

1 The inscription runs as follows (see Archaeologia, l, 70) : — Odda Dux 
jussit hanc regiam aulam construi atque dedicari in honore S. Trinitatis pro 
anima germani sui ^Ifrici que de hoc loco asupta. Ealdredus vero Eps qui 
eandem dedicavit 1 1 Idibus Apl. xiiii aute annos regni Eadwardi Regis 
Angloru. The fabric would be properly described as a votive or memorial 
oratory. Its size and relation to the main church of the place are suitably 
expressed by the term * chapel.' 










proportion to its width, while the view indicates that the 
height of the nave walls is similarly marked. The gable 
is also sharply pointed. The exterior view from the south- 
east, Fig. 62y will enable a judgment to be formed of 
the technique. The walls which are 2 ft. 3 in. or 4 in. 
thick 1 are constructed for the most part of squared stones of 
ample size, many of which show by their tooling or other 
marks upon them that they are Roman stones brought from a 
neighbouring station of the legions.^ Blocks of a specially 
large size are used for the quoins and some of these are i ft. 
6 in, to 2 feet in height by 3 ft. to 4 ft. in length. Notice 
should be taken of the manner in which they are laid. They 
are set up on edge and extend like slabs along the wall alter- 
nately north and south and east and west as do the quoin- 
stones at Stow illustrated in Fig. 23y ante, p. 87. There is no 
trace of the technique of long-and-short work. On the other 
hand this meets us as soon as we enter the building (Fig. 
64), in the imposing and characteristic feature of the chancel 
arch. This measures 1 5 ft, in height by a width of 
5 ft. 3 in. and is constructed of large stones that all go 
through the whole thickness of the wall and are carefully 
squared or cut to the wedge-shaped voussoir form. The 
jambs are fashioned like those at Heysham or Somerford 
Keynes of slabs alternately upright and flat. The imposts are 
chamfered and the impost-stone on the south is thicker than 
the other, so that a portion of the jamb is cut in the same stone 
below the chamfer. A glance at the Roman door jamb in 
Fig. I, ante, p. 4, will reveal the same peculiarity. 

Next in order may be noticed the two ancient doorways 
surmounted by flat lintels. They are both in the north walls, 
one, now built up, in that of the chancel and the other, shown 

1 The walls are said to batter as they rise, i.e. to be thinner at the top. 
They are certainly not quite plumb, and only careful measurement, which now 
would be hardly practicable, could decide if they really taper. 

^Probably from the Roman station ofVinovium now Binchester. 


] I 

in Figs. 65 and 66, m the wall of the nave. The jambs, it 
will be noticed, are constructed like those of the chancel arch. 
There is however the farther peculiarity which we meet with 

Fig. 64. — Interior view of Escomb Church, Durham. 

here for the first time, that the jambs are slightly inclined so 
that the doorway measures 3 ft. in width above, but 3 ft. 3 in. 
below. This north opening it must be observed is rebated for 
a wooden door, which was kept closed by the common device of 
the wooden bar that played into a recess in the stonework of 





the jamb. On the outside we find the curious arrangement 
that the flat lintel and the jambs are mortised into each other 
after a fashion that reminds us of the Roman gateway shown 
in Fig. 2, ante, p. 5. A similar peculiarity will meet us at 
Britford near Salisbury. Besides these two north doorways 
the present south doorway of the nave is an original opening 
though it has been altered in modern times. Part of the 

If — rg^ 

--••wni -"v^ol 

^.-^J-KjI/l'iV^v-'V- ■..ii>^"^'>=acrs^. 

Fig, 65. — North doorway of nave, Fig. 66. — North doorway of nave, Escomb, 

Escomb, interior view. 

exterior view. 

east jamb is ancient. It must be remarked that the north 
and south doors are not opposite each other, though this 
symmetrical arrangement is commonly found. 

The small original lights on both sides of the nave are 
noteworthy, Figs. 67-8. The two on the north are flat-headed 
but the southern pair have round heads cut out of single 
stones, or rather out of two single stones set one behind the 
other to make up the thickness of the wall. The innermost 
lintel of the south-east window is fully 7 ft. in length. Both 
windows are internally splayed and have markedly sloping 
jambs. The aperture is in each case on the outer face of 
the wall and measures some 2 ft. 8 in. by i ft. 5 in. The 
internal opening is about 5 ft. high by a mean of about 



2 ft. 7 In. in width. The groove for a shutter is visible on 
the jambs of the round-headed light. No original openings are 
preserved in the chancel but there is one high up in the 
west gable of the nave. 

There are remains of old, possibly original, plastering on 
parts of the walling of the church, and a bit of pebble 
flooring in the north-west corner. On the exterior, on 

Fig. 67. — Round-headed window internally 
splayed, south side of nave, Escomb. 

Fig. 68. — Square-headed window, 
north side of nave, Escomb. 

the south wall of the nave is a mutilated Saxon sundial 
with a curious monster, half serpent and half fish, curling 
round it. 

The modern history of the wonderful little church is a 
curious one. It had apparently been always in use till 1863 
when a new church was built a little way off on the higher 
ground, and the old building, the value of which no one 
seems to have suspected, was allowed to fall into decay. 
In 1879 attention was for the first time directed to it from 
the archaeological standpoint and its antiquity and interest 
were at once recognized. It was properly repaired and is 
still in use for occasional services for the population of 
what is now a mining village. 


HI. The Oratory with Apsidal Presbytery. 

The next type consists in a single oblong chamber with 
a sanctuary not square but apsidal. Nearly all the 
examples of this aisleless apsidal type that are known to exist 
are in Kent, but it does not follow that they were not extant 
at one time in other parts of the country. The example 
which we will take first to illustrate the type is a building 
with which some historical difficulties are connected. It is 
near Bradwell in Essex and is known as St. Peter-on-the- 
Wall, from the fact that it is actually built on the site 
of the principal gateway of the Roman fortress of 
Othona (see the map of Roman Britain, Vol. i. Fig. 5, 
p. 52). This was called in Bede's time Ythancaester, and 
here as the historian tells us Cedd, the apostle of the 
East Saxons, about 653 a.d. built for himself a mission 
station. 1 There is now to be seen on the site a desecrated 
structure that has many of the characteristics of an Early 
Saxon church, though it is not exactly of the kind that 
Cedd would have built, nor is it placed where at Cedd's 
time such a church would naturally be located. The form is 
apsidal, and the apse belongs to Roman building tradi- 
tions, not to those of the Celtic Church of which Cedd 
was an alumnus ; moreover the custom of the time was 
to place Christian churches and mission stations within 
the protecting lines of Roman forts, a course adopted by 
Fursa in East Anglia,'- and not to level the walls of these 
and build on the top of them, as has been done at St. 
Peter-on-the-Wall. Hence it has not been without con- 

^ Vol. I, pp. 172 and 201. Othona and the building under consideration 
form part of the subject of a paper by Th. Lewin in Archaeologiay 
XLi, 421 f See also Arckaeologica/ Journal, lviii. 

2 Vol. I, p. 201. 


siderable misgiving that the writer has yielded to the 
impression made upon him by the interesting monument 
and has included it in this survey. 

The description of the building which is now used as 
a barn is easy. The plan shows an oblong cella laid 
out with far more accuracy than is generally the 
case with Saxon work and measuring internally 49 ft. 
8 in. by 21 ft. 8 in. This was terminated to the east 
by a semicircular apse now destroyed. The walls 2 ft. 4 
in. in thickness are chiefly composed of Roman materials, 
both brick and stone, and exhibit the uncommon feature 
of buttresses, of which four strengthen the two western 
quoins and four others abut the north and south walls. 
One on the south side is well preserved and is 2 ft. on 
the face with a projection of i ft. 10 in. At a height 
of some 12 ft. they end with sloping heads. These 
buttresses are in part at any rate in bond with the wall, 
as can be seen, e.g., in the fragment of the easternmost 
one on the south side. In one place a large stone is 
partly in the wall and partly in a buttress and is cut to 
the angle formed by the two faces. There was a west 
door and apparently a porch in front of this. Five 
windows, though considerably battered, have left their 
traces, two in each of the long walls and one high up 
in the west gable. They are internally splayed from an 
aperture of about 3 ft. in width to an inner spread of 
about 5 ft. 

In place of the normal ' Arch of Triumph ' which we 
should expect to find opening towards the apse there were 
smaller arches, apparently three in number, the northern- 
most of which is preserved for a space covered by 
twenty two of the Roman bricks of which it is composed. 
The arch is set back an inch or two on its jamb. This 
peculiarity, of an arcade in place of a single arch between 
nave and presbytery, will be noticed later on. 


On the eastern side of this arcade the line of the nave 
wall is extended on the south for about 4 in. where, in 
the lower part, it ends with a smooth face that seems to 
indicate a doorway. Mr. Lewin stated that there was 
here on the north a sacristry or vestry, of which the 
foundations were found by excavation.^ On the south 
the corresponding wall runs in a straight line for 4 ft. or 
so, so that the curve of the apse did not begin for some 
little distance beyond the arcade. This ' stilting ' of the 
semicircle of the apse, for so it may be termed, is found in 
many Early Christian churches, and is specially well repre- 
sented in those of North Africa and Syria." 

There are so many features about the structure that 
betoken an early date that it is impossible not to include 
it in this survey. The better known Kentish examples 
may now receive attention. 

The old church at Lyminge in Kent, already referred 
to (Vol. I, p. 279), the foundations of which are to be seen 
immediately to the south of the present parish church, 
was of this plan and measured 32 ft. by 17 ft. 3 in. in 
the nave, with a semicircular apse 14 ft. 6 in. in diameter ; 
and of this plan too was the early Saxon church at Rochester, 

^ Archacologia, loc. cit. p. 448. 

^As this is the first introduction in this survey of the apsidal termination, it 
may be noted that existing remains or traces of ten Saxon apses show 
the following variations in plan : Lyminge, Reculver, St. Peter-on-the-Wall, 
Deerhurst, and Worth are semicircular ; Rochester, Lindisfarne, and probably 
St. Pancras, semi-elliptical ; Brixworth is rounded internally but on the 
exterior polygonal ; Wing polygonal inside and out. The form in which the 
apse is made up square on the exterior, a variation not uncommon in the Early 
Christian churches of North Africa and of the East, does not occur in 
extant Saxon work. Some excavations begun in 1901 at Much Wenlock, 
Shropshire, on the site of a conventual settlement of the seventh century, 
seemed to show indications of a building of this form, but they have not 
yet been carried far enough to settle the question. The use of the apse in 
general in our early churches will form the subject of a paragraph in a 
later chapter. (See postea, p. 279 f) 



the outline of which has recently been recovered by 
excavation (Fig. 69).^ 

Rochester is fully discussed and illustrated by the Rev. 
Grevile M. Livett and Mr. St. John Hope in Archaeologia 
Cantiana^ Vols, xviri and xxiii, and the accounts 
make it clear that the remains represent a Saxon church 
of repute, which may very well have been the first bishop's 
church on the site, dating from the year 604 a.d. The 
nave of this church measures 42 ft. by 28 ft. 6 in., the 

Fig. 69. — Plan of early Saxon church at Rochester. 

apse has a diameter of 24 ft. 6 in. by a depth of 19 ft., 
exhibiting thus the form of a half ellipse rather than a 

The most interesting, because possibly the earliest, example 
of the plan under consideration occurs in connection with a 
building the architectural history of which is obscure, but 
which is quite the most famous parish church in the whole 
of England. The reference is to the church of St. Martin 
by Canterbury, mentioned by name by Bede and associated 
with the first introduction of Christianity into Saxon England. 

^ The writer is indebted to Mr. St. John Hope lor permission to reproduce 
this plan. The dimensions given for Lyminge are taken from Mr. Hope's 



It consists now in a square western tower, nave, and long 
square-ended chancel, and the nave and the western part of 
the chancel are clearly of pre-Conquest date. There is some 
evidence, discussed in detail in a recent work on the church 
bv Canon Routledge,^ that this western part of the chancel 
was the original church. It is supposed to have terminated 
towards the east in an apse and to have extended westwards 
into the part which is now the nave so as to form in this way 
a plain round-ended chapel. Later additions and alterations 
brought the church to its present shape as shown in the 

N4 AR ^''•>i»-nl Ci*>i^~tf^' 



P i ^ \w |' v ^< UWJlW >< ^ l <>llUWlWtV ' H * A ' »<WWfc ' ^ »T 'W W< ''J>. ti ^Vl>Lm« 

ejira« 510N5 


I l-^l ;.J-Ui 

Fig. 70. — Plan of St. Martin, Canterbury, showing supposed original form 
of the church in the present chancel. (Special scale.) 

plan, Fig. 70. In this plan, the work of the Rev. Grevile 
M. Livett, which is here copied with the kind permission 
of Canon Routledge from his book, the supposed original 
nucleus of the structure is sufficiently indicated, and an 
examination of the building in its existing condition demon- 
strates the high antiquity of the western part of the 

The present south wall of this, 2 ft. 2 in. thick, is com- 
posed almost entirely of Roman brick laid with fair regularity. 
About half-way along the present extended chancel wall 
there is a projecting buttress of the same material (now a 
good deal modernized) and immediately to the east of this 

^ T/ie Church of St. Martin, Canterbury. London, 1898. 


there is a straight joint in the brickwork showing where the 
orio;inal chapel ended. Of the two doorways, now blocked, 
in this wall, the round-headed one is a later insertion and 
has sloping jambs with an arch head wider than the jambs. 
The flat-headed one at the south-west corner is probably 
original, and may have given access to a small lateral chapel. 
The eastern part of the present chancel is much later. On 
account of the disturbances caused by interments within the 
chancel it has been impossible to ascertain decisively whether 
or not the original chancel really ended apsidally, though this 
is rendered highly probable by the treatment of the south-east 
corner where there is the buttress on the south but none on 
the east, though there would have been one here also had 
the chancel ended in a straight eastern wall.^ 

The nave of St. Martin is also Saxon, or at any rate pre- 
Conquest," but the walls of it are of ruder construction than 
those of the chancel, though they exhibit very distinct sur- 
vivals of Roman technique. One of these consists in a 
patch of red plastering, visible near the little piscina at the 
south-east corner of the nave, and it appears that there are 
considerable pieces of this material behind the woodwork 
of the pews. This plastering, made with pounded brick, is 
hard and of good quality, and might in itself be termed 
a specimen of the opus signinum of the Romans. Another 
significant feature was only displayed to view when the west 
wall of the nave was recently denuded of its plaster. Here 
have come to light marks of three openings which had been 
walled up ; one, in the centre, is a large arched opening, 
about 7 ft. 6 in. in width, with its crown about 17 or 18 ft. 
from the floor, the object of which is problematical, and the 

^ The example of St. Pancras presently to be noticed shows that this 
would have been the case. 

^The question whether the original work, in St. Martin, and in others 
of the earliest churches in Kent and elsewhere, is Saxon or pre-Saxon must 
for the present be reserved. 


other two are splayed windows, about 4 ft. 6 in. high, the outer 
openings of which are permanently concealed by the walls of 
the mediaeval tower. These windows have jambs composed 
mainly of blocks of chalk, and round heads turned in Roman 
bricks and voussoirs of Kentish rag buried in abundant mortar. 
The peculiarity here is that this mortar is of the pink kind, 
composed in part of pounded brick, while that of the walling 
generally is white. There would be no special significance in 
this but for the fact that precisely the same peculiarity occurs 
in the original arched openings in the already mentioned Roman 
Pharos at Dover Castle, which are turned in mortar mingled 
with crushed tiles, while in the fabric generally the mortar 
is white. It is enough to record these facts, the historical 
purport of which cannot here be discussed. 

One of the most instructive of our apsidal churches is that 
of St. Pancras at Canterbury, the remains of which, consisting 
(save in one portion) only in the lower courses of the walls, 
are to be seen in a field to the east of St. Augustine. They 
have recently been laid bare in connection with important 
excavations now proceeding near the site, and are made the 
subject of a paper by Mr. St. John Hope in Archaeologia 

The ground plan. Fig. 71, will be seen at once to promise 
features of novelty and interest. The main body of the 
building is an oblong measuring internally about 42 ft. 8 in. 
by 26 ft. 9 in., and it will be observed that the western 
quoins are buttressed, each wall being carried out for the 
purpose. Two similar buttresses or pilasters flanked the original 
wide entrance doorway at the western end, that measures 
7 ft. 6 in. Notice should be taken of these buttresses as 
the feature is very rare in pre-Conquest work. They are 
of the same width as the thickness of the walls which is 
never more than about i ft. 9 in. and project i ft. 2 in. 
They occur on each side of the angles to which they give 

^ Vol. XXV. 



strength (cf. St. Martin, ante, p. 121) and are in bond with 
the fabric of the walls. At the eastern end the oblong cella 



was terminated by a screen of columns or a wall, with a wide 
archway about 9 ft. in span in the centre, giving admission 
to a presbytery or chancel the walls of which, enclosing a 


space rather less in width than the nave or cella, were carried 
on parallel lines for some lo ft. and then came together in 
the form of a wide apse, probably of semi-elliptical plan, that 
terminated the whole interior. A buttress marked on the 
exterior the point where the curve of the apse began. Only 
the portions marked in black on the plan have actually been 
recovered and the rest is more or less conjectural. 

The material is Roman brick, re-used, with the original 
Roman mortar in many cases still adhering to the surface, 
and these bricks or fragments of bricks are laid in two 
kinds of mortar, one yellow which was used in the begin- 
ning of the work, and the other white with fragments 
of sea-shells in it, that was employed later, but still before 
the fabric was more than half completed. The bricks are 
sometimes of triangular form and laid with the points inward 
according to a familiar Roman fashion. 

Furthermore, the building has three adjuncts in the form 
of a western porch and two lateral chapels to which attention 
must be paid. Each of these features possesses in Saxon 
architecture a special importance over and above its own form 
and use. The western porch, as will be seen later on, is 
closely connected with the conspicuous feature of the 
western tower, while the side-chapel leads us on to the transept 
and is associated in this way with the development of the 
cruciform plan. At St. Pancras the walls of these three adjuncts, 
as is shown on the plan, are not in bond with those of the 
main building, and they are constructed with the white 
mortar that signifies a slight posteriority in date. Such 
at any rate is the case at the lowest part of the walls, as for 
example where the walls of the western porch are built up 
against the pilasters that flank the western door. The north 
wall of this porch is however preserved to a considerable 
elevation, and it can be seen that, after the height of about 
3 ft. has been reached, the joint between the porch wall and 
the pilaster disappears and both are carried up together with no 


division. This points to the explanation that the porch was an 
afterthoug-ht but that it was taken in hand before the work 
on the main building had proceeded far. The same is prob- 
ably the case with the two lateral chapels. They are not, 
so far as can be seen, in bond with the main walls, 
but, as in all the rest of the building save the north 
wall of the porch, only a few of the lower courses of 
the walling are preserved. Higher up they may have 
been bonded in like the western adjunct. Of these lateral 
chapels only that on the south has been preserved, but there 
are indications which show that there was once a corresponding 
structure on the north. The southern chapel contains against 
its eastern wall the relics of a later mediaeval altar that is of 
special interest for us as it appears to be referred to bv a 
writer of the fourteenth century, who tells to us about the 
altar and the building generally a noteworthy but perhaps 
hardly credible story. 

The writer in question is one William Thorn, a monk of 
the monastery of St. Augustine, within the precincts of which 
the chapel we are dealing with is situated. He tells us that at 
this place King i^thelberht before his conversion possessed a 
temple or idol house, and that Augustine purged this temple 
from its impurities and changed it into a church dedicating it 
in the name of St. Pancras.^ ' There is still an altar,' Thorn 
goes on, ' in the south porch ^ of the same church at which the 
same Augustine was wont to celebrate . . . ' and it is evident 
from what he goes on to tell us that the chapel and altar had a 
traditional sanctity in later mediaeval days which gives them a 
certain importance in our eyes. Thorn's statement that the 
building was originally, or at any rate at one time, a pagan 
temple is of more immediate moment. 

This same statement has been made in earlier and in later 
days about not a few of the oldest-looking of our Saxon 
churches especially in Kent. It has never in the case of any 

^ For the reason of this dedication, see Vol i, p. 279. ' Postca, p. 129. 


one been proved, and the tendency now-a-days is to discredit 
any suggestions of the kind. It is quite possible that the 
notion to which Thorn gave currency was a mediaeval guess, 
founded on the fact that antique Roman columns, or portions 
of these, were to be seen as part of the fabric at the eastern 
end of the nave. 

The use of these columns, and the fact that the apse does 
not begin immediately beyond them but only after an inter- 
mediate rectangular space, are peculiarities that amount to the 
creation of a new type and this we will accordingly consider 
under a fresh heading. 

IV. The Oratory with Presbyterial Space Screened Off 

Before the Apse. 

In the simplest form of the Early Christian church, that of 
an oblong interior terminated by an apse (chap, i, p. 14) the 
apse opens immediately at the end of the rectangular cella 
through an arch that takes in its total width in one span. In 
the chapter just referred to it was pointed out that in more 
advanced churches various arrangements were adopted for 
securing additional room in this presbyterial portion of the 
interior, by interposing variously shaped and contrived spaces 
between the nave proper and the actual altar-house. This was 
the beginning of the development of the choir, which in later 
mediaeval times assumes such large proportions in relation to 
the rest of the edifice, England is conspicuous for the extent 
to which in her churches this development was carried, and 
the choir of Canterbury cathedral, in its present stupendous 
extension, represents the final outcome of a process that in 
the same city we can see beginning at St. Pancras. Among 
English pre-Conquest churches that of Brixworth best re- 
presents this extension of presbyterial space, and any remarks 
that have to be made on it, with the necessary indication of 
foreign parallels, will be reserved until Brixworth is reached. 


The other peculiarity noticed above, that it is not a single arch 
but an arcade or a screen which is interposed between the nave 
and presbytery, may have here a word. This implies the substi- 
tution of several narrower arches for one wide one, and this was 
rather in accordance with English tendencies. It is a character- 
istic of English architecture through the whole mediaeval period 
that even in edifices of great importance vault construction is 
rather avoided than favoured. There are it is true great English 
achievements in vaulting of which Durham is the chief, but in 
the main what has just been said holds good. As a fact, the 
builders of mediaeval Scotland seem to have been far more at 
home in vault construction than their brethren south of the 
Tweed. Among all the numerous specimens of pre-Conquest 
architecture in England there is only one that exhibits a vault 
in any other position than in a crypt. None of the existing 
Saxon apses has retained a vault if any ever possessed such a 
finish, and it is only in the porch at Monkwearmouth that an 
above-ground vault is to be seen. In arch construction there 
is the same deficiency, for though there are well-constructed 
Saxon arches, such as the tower arch at Barnack, 13 ft. wide, 
and the chancel arch at Worth, Sussex, which is over 14 ft. 
in span, yet as a rule the openings of Saxon doorways and 
chancel and tower arches tend to narrowness, and at Bradford- 
on-Avon the chancel arch is only 3 ft. 6 in.^ in width, and 
one of the principal doorways only a little over 2 ft., while 
we find again and again examples of faultily cut voussoirs,- 
which show that the elementary principle of the radiating 
joint was by no means universally apprehended among Saxon 

^ For purposes of comparison it may be noticed that some of the pre- 
Norman Scottish churches, old in type if not always in actual date, have very 
narrow chancel arches (Anderson, Scotland in Early Chr'utian Times, i, 60 f ) 
and so had an early oratory now almost buried in the sand near Gwithian 
in Cornwall. On the other hand, many of the early nave and chancel 
churches in Ireland have wide arches. See Fig. 15, ante, p. 29. 

2 e.g., the tower arch at Bosham, Sussex, see Fig. 94, postea, p. 170. 


It is not therefore surprising to find the noble arches of 
ample height and span which terminate the naves of the normal 
Early Christian churches of the Continent replaced at times 
among ourselves by arcades, generally of three openings of 
smaller span. This arrangement occurs in the known Saxon 
churches of Reculver and Brixworth ; also at St. Peter-on-the- 
Wall. Designed at St. Pancras, it is conjectured with some 
approach to certainty at Rochester, with perhaps Lyminge. 
In later times there is a sort of make-shift reproduction of 
this early arrangement in squint-like apertures pierced on each 
side of a narrow chancel arch. One instance occurs in the 
fine Saxon church at Bracebridge by Lincoln. It is doubtful 
however whether any of these are really ancient. Some are 
certainly quite modern. A parallel in later work to the 
early Saxon chancel arcade is however to be found in the 
church at Westwell, near Ashford, in Kent, where a beautiful 
Early English arcade of three trefoil arches divides nave from 

The most interesting of the Saxon examples is that which 
existed at Reculver and this will be noticed on a later page. 
With the instance now before us at St. Pancras certain diffi- 
culties are connected, the full discussion of which would occupy 
space that cannot be spared. There is evidence here of the 
existence of four antique Roman shafts which might well 
have been the supports of an arcade between the nave 
and the presbytery, and the base and lower portion of 
one of these shafts is still in situ at the southern end of the 
original arcade, where it is indicated on the plan. The 
southern wall of the nave is however returned northwards at 
its eastern extremity, and brickwork in bond with this nave 
wall, and so belonging to the earliest stage of the building, fills 
in the space between the wall and the shaft, the bricks where 
they impinge on it being neatly cut to shape. Furthermore, 
another portion of wall representing the second, but still nearly 
contemporary, stage of the building, fills in the whole space 


between this southernmost shaft and the next one in the row. 
Between the latter and the third shaft, on the other side of the 
central line of the building, there was an open archway of 
about 9 ft. in span, and this was the really effective chancel 
arch. The arcade of columns for some reason or another was 
apparently never an effective feature but was built up before 
the church was completed. 

V. The Porch, Western or Lateral. 
VI. The Lateral Chapel. 

St. Pancras has already introduced us to the western porch 
and to lateral structures which as they have no external opening 
are rather chapels than porches. Both of these would have 
been described in mediaeval times by the word ' porticus ' and 
on the uses of this term a sentence or two may usefully be said. 

In classical Latin porticus as connected with ' porta ' meant 
a vestibule, but as one of the most familiar architectural em- 
bellishments of a portal was a canopy supported on pillars, 
the word gradually acquired the sense of a colonnade. Porticus 
may therefore have the legitimate meanings of any columned 
structure or anything in the shape of a vestibule, and we 
find that in addition to these senses the word was used to 
denote any side space or adjunct opening into the main body 
of a building, though not actually a vestibule. Ducange, 
sub voce, quotes a mediaeval writer who even uses porticus 
for the sanctuary or altar end of the church as a whole. 
The lateral chapels at St. Pancras would be called porticus, 
while the western vestibule should in strictness have a 
qualifying word and be termed ' porticus ingressus.' When 
Dunstan at Glastonbury wished to make the width of a certain 
older church correspond to its length he is said by William of 
Malmesbury to have added 'alas vel porticus'^ and this 

1 Memorials of St. Dunstan, Rolls Series, No. 63, p. 271. 


obviously implies side aisles, which would come within the 
scope of the word owing to the use of a row of columns in the 
nave arcade. Effmann says that porticus is also used for the 
concentric aisle round a central circle or octagon, as at San Vitale, 
Ravenna ; or Aachen.^ On the other hand when Bede tells us of 
the burial of King i^thelberht that he was laid ' in porticu Sancti 
Martini,'- in the church of St. Peter and St. Paul (St. Augustine) 
he cannot be speaking of an ordinary side aisle for this would 
not be specially dedicated to a saint. He may be referring to 
a lateral chapel like those at St. Pancras, but there is the possi- 
bility that this porticus was one of the side arms of a Greek 
cross, a plan orthodox, as we have seen, for a sepulchral 
church. Such a use for porticus we find in Prior Richard of 
Hexham who tells us that Wilfrid's ' central ' church of 
St. Mary was ' a quatuor partibus totidem porticus habens,' ^ 
The choir and transepts of a cruciform church on a Latin plan 
are called porticus in a passage quoted postea, p. 242. 

With this note on the very wide and varied use of the 
word porticus in mediaeval literature, we may go on to con- 
sider the porticus in the sense of porch or vestibule rather than 
in that of side chapel. 

This feature as we have seen * plays an important part in 
the economy of the old English church, and some of its 
religious and social uses have been noticed in the foregoing 
volume. It existed as an adjunct both of the churches with 
square-ended chancels and of the apsidal ones, and it some- 
times appeared at the western end (St. Pancras, Monkwear- 
mouth, Corbridge), sometimes on the south (Canterbury 
cathedral, Bishopstone), or again sometimes on the north side 
(Bradford-on-Avon). The shape and position of the St. 
Pancras western porch can be seen on the plan. It was 
entered from the west by an archway the imposts of which 
can be discerned on the existing north-western jamb, and 

^ Die Karol. Otton. Bauten zu Werden, p. 39. ^H.E. ii, 5. 
^Twisden, Decent Scriptores, col. 290. * Vol. i, p. 369. 




which was about 1 1 ft. In height by a width of about 6 ft. 6 in. 
There are no signs that it was closed by a door. The other 
early western porches at Monkwearmouth and Corbridge in 
the north, have had towers reared over them and so have 
lost their ancient character. 
They will be noticed later on. 
Not many Saxon lateral porches 
actually survive but more modern 
porches have often been con- 
structed in front of original 
Saxon doors, as is the case at 
Breamore, Hants ; Dagling- 
worth, Gloucestershire, etc. Of 
existing lateral porches the most 
imposing is to be found at 
Bishopstone, near Seaford, in 
Sussex. The plan given in 
Fig. 72 shows the porch and 
western part of the nave which 
are of Saxon work. The square 
western tower and the eastern portion of the building, as 
seen in the general view, Fig. 73, are Norman. 
/The porch at Bishopstone has a little Norman secondary 
porch applied in front of its own doorway, above which 
is a Saxon sun-dial. The height of the gable is 21 ft. 
There is some massive quoining in large squared stones at 
the outer angles, one stone being 4 ft. 3 in. high by i ft. 6 in. 
by I ft. The structure provides an interior space of 12 ft. 
5 in. north to south by 9 ft. 2 in. east to west. The door 
into the church is not in the centre of the porch, but on 
the western side so as to leave the eastern part of the porch 
interior free, and this is a fact of some significance, as will 
be seen whpn we turn our attention to the porch or porches 
at the well-known Saxon church at Bradford-on-Avon in 
Wiltshire, of which the plan and section are given in Fig. 74. 

Fig. 72. — Plan of southern porch 
and western end of nave, 
Bishopstone, Sussex. 



This remarkable little monument is of the nave and square- 
ended chancel type but has been reserved for this place on 
account of its lateral porches. A few general words upon it 
may prove of interest to the reader. 


Fig. 74. — Plan, and section of nave, of the Saxon Church at Bradford-on- 

Avon, Wilts. 

William of Malmesbury in his life of Aldhelm states that 
the West-Saxon bishop was generally supposed to have built 
a monastery at Bradford,^ and adds, ' to this day at that 
place there exists a little church which he is said to have 
made in honour of the most blessed St. Laurence.' "^ Writing 
in 1858 the Rev. Canon Jones, then vicar of Bradford-on- 
Avon, stated that the site of Aldhelm's monastery ' was most 
probably near the north east end of the present (parish) church, 
a spot of ground there still bearing the name of the Abbey 

^ Necnon et apud Bradeford tertium ab eo monasterium instructum crebra 
serit opinio. (The other two were at Malmesbury and Frome, Somerset.) 

2 Et est ad hunc diem eo loci ecclesiola quam ad nomen beatissimi Laurentii 
fecisse predicatur. Gesta PontificuTtiy Rolls Series, No. 52, p. 346. 









yard.'^ On the site thus indicated there existed a confused 
complexus of buildings of various dates, and from the midst 
of these the perspicacity and zeal of Canon Jones succeeded 
in extricating the highly interesting little Saxon building 
which in its restored condition Fig. 75 presents to the reader. 
When the building was taken in hand for restoration the 
chancel was divided into two stories and used as a cottage 
the chimney of which went up where the chancel arch had 
formerly been. The nave, also with an intermediate floor, 
was a school, and the schoolmaster's residence was in a 
house built up against it on the south. A woollen manufac- 
tory abutted upon it on the north and the western part had 
been modernized, so that everybody supposed it of the 
eighteenth century. About 1870 it was cleared and restored, 
but as needs hardly to be said the masonry has been so 
knocked about, patched, and renewed, that it is only in 
parts that we find any large surface of original work. 
This when it can be examined shows masonry of largish 
stones, running to some 2 ft. in height and width by 
I ft. in thickness, of the excellent local material the use of 
which gives the whole town of Bradford so handsome 
an appearance.2 The stones are well cut though not 
accurately squared and a large number of the joints are 
not vertical or horizontal but sloping. The actual jointing 

^ Wiltshire Archaeological and Natural History Magazine, v, 1 2. 

2 There is a notice of Bradford-on-Avon in Freeman's English Tozvns and 
Districts, Lond. 1883. It is unfortunate that local patriotism has not availed 
to secure the preparation of a proper guide to the Saxon church that is so 
much visited. The guide now sold is a libel on the memory of its titular 
author. Dated 1892 it spealcs of a time before 1857 as 'about twenty years 
ago ' ; it talks of the modern house as still abutting on the southern side of the 
nave, though this has been removed years ago ; and worst of all presents the 
visitor with a plan in which there is an obvious error in the placing of the 
north door into the porch. On page I the church is claimed to be of the 
early part of the eighth century, but by pp. 21, 23, it has apparently been 
transferred to the tenth. 


of the stones is however remarkably fine, and excited the 
admiration of such an expert in masonry as the late J. T. 
Irvine, who left among his papers a valuable set of drawings 
and notes about the church, made during the restoration 
from 1869 to 1874. 

The walls, which have a mean thickness of 2 ft. 5 in. exclu- 
sive of plinths and pilasters, rise from a plain square plinth now 
owing to the accumulation of the soil only visible on the south 
and east. The quoins exhibit no special treatment, except what 
they receive in connection with the general scheme of external 
enrichment which is one of the peculiarities of the structure. 
This consists in a series of pilasters in the lower story 
of the elevation and arcading above, the two being separated 
by the projection of a horizontal string course. The pilasters 
embrace the angles, and the various wall spaces are divided 
below by intermediate pilasters, some of which rise with a 
step-like base from a shallow plinth distinct from the lower 
projecting plinth just mentioned. The short pilasters of the 
arcading above have trapezoidal bases occasionally stepped, 
and plain trapezoidal caps. On the east face of the chancel 
and in the east gable of the nave and north gable of the 
porch they are reeded. The scheme of decoration seems to 
have been carried out by cutting into the face of the finished 
wall to a shallow depth, the jointing of the stones being in 
parts of the work entirely ignored, so that at first sight the 
effect is that of incised enrichment rather than of architectural 
ornament proper, which is more closely connected with 

To say this however does not imply that the decoration is 
out of all relation to the fabric, or that, as Sir John Henry 
Parker suggested, it might have been added at a later period to 
an already existing unadorned structure. A careful examina- 
tion of the work, especially in regard to the planes of its various 
surfaces, shows that the enrichment was planned when the stones 
of the walling were laid, and is necessarily contemporary with 



















the fabric. Fig. 76 shows a portion of it on the north side 
of the chancel drawn to scale/ with indications of the relations 
as regards projection of the different faces. The string course, 
shown in the general view (Fig. 75) as dividing the wall 
horizontally at about two thirds of its height, is formed all 


Fic. 76. — Pilasters, arcading, etc., on the exterior of Bradford-on-Avon, with 
section of chancel wall. Scale ^ of nature. The original wall face is 
represented by the face of the upper and lower pilasters and the part 
above the arcading. 

along in a single course of stones 6i in. high, and always 
projected about i in. from the main face of the wall. The 
trapezoidal bases of the pilasters of the arcade above the 
string course, with the parts between them, are also formed in 
a single course of stones, and the same is the case with the 
capitals above the pilasters and the parts between them. The 

^ Some of the smaller features in the drawing have been taken from other 
parts of the building, in order that a conspectus of the peculiarities of the 
work may be presented. 


height of these courses is much less than the average height of 
the wall stones generally, and whereas the wall stones vary in 
size in the most irregular manner, these particular courses run 
practically without a break all round the building. 

P'urthermore, the face of the trapezoidal caps is in projection 
over the original main face of the wall, just as is the face ot 
the string course below, and before the arcading was incised 
this course wherein the caps are cut must have stood out like 
the string course, though with somewhat less projection. 
Whether this was also the case with the course in which the 
bases are cut is in the present condition of the work not easy 
to decide. The pilasters lastly, though generally cut out of 
wider stones, are in some cases, as on the east face of the 
chancel, in stones of just the width required, while in almost 
every case the height of them, about 2 ft., is in a single stone, 
thus showing that the pilasters, like the caps and bases, were 
prepared for in the structure of the wall. These observations, 
which can be made on the building in its existing condition, are 
borne out by the results of an examination of the actual fabric 
of the east wall of the chancel made during the restorations, 
and embodied in one of Mr. Irvine's invaluable drawings 
reproduced at the right hand side of Fig. 76, 

It is clear therefore that the enrichment, though of the 
incised kind, is in close connection with the structure with 
which it must necessarily be coeval. The fact thus established 
has a bearing on the debated question of the date of the 
building. Some idea of the probable date of the decorative 
arcading, and hence of the whole structure, may be gained by 
comparing it with a feature bearing a remarkable resemblance 
to it in a church in a distant county. The reference is to a 
scheme of shallow arcading which occurs round the interior 
of the nave of the Saxon church of Dunham Magna, Norfolk 
(Fig. 77). The scale is rather larger than at Bradford. 

We enter now the little church through the doorway into the 
north porch. This is extremely narrow and the jambs slope con- 




siderably, being 2 ft. i in. apart under the imposts but about 2 ft. 
4 in. below, and it is much to the west of the middle of the 
wall. The reason for this doubtless is that an altar was originally- 
placed against the eastern wall within the porch, which was in 
this way turned into a lateral chapel, and the same surmise may 
be made about the porch at Bishopstone. It has been a subject 
for discussion whether or not there was a corresponding porch 
or chapel on the southern side. Conclusive evidence that this 
was the case came to 
lip-ht in the course of 
the restoration, and 
fully satisfied Mr. J. 
T. Irvine who had been 
at first disposed to 
doubt it. Now that 
the southern side of 
the nave has been set 
free and cleaned, the 
mark of the old porch 
is plainly to be seen 
upon it, and is quite 
distinct from the marks of the house that till recently was built 
up against this side of the little structure. If we replace on the 
plan this second porch we shall find that the two porches 
measured together offer an interior space equal to nearly two- 
thirds the area of the nave. Lateral porches of this size, especially 
when used also as chapels, become something like transepts, and 
as we shall see later on have an importance in connection with 
the general development of ground plans. 

On entering the body of the church we are struck all at once 
by the great proportionate height of the side walls, of which 
Saxon peculiarity Bradford presents an extreme example. There 
is a plain plinth all round the walls of about 2 in. projection. 
Leaving this out of account, the nave measures nearly 
25 ft. in length by 13 ft. 8 in. in width, but the height 

Fig. 77. — Arcading in the interior of the nave 
of Dunham Magna, Norfolk. 


to the top of the walls is more than 25 ft., so that the 
interior is as high as it is long and nearly twice as 
high as it is wide. The section appended to the plan, 
Fig 74, shows this peculiarity. In connection with this it 
is important to know that there is no trace that the church or 
any part of it ever had two floors. This suggestion, which 
had been made, was decisively negatived by the results of 
Mr. Irvine's technical examination of the structure when 
under restoration. The chancel, measuring 13 ft. 4 in. in 
length by a mean of 10 ft. 2 in, in width, is entered from the 
nave through an archway 3 ft. 6 in. wide, that is for the most 
part a restoration. It is the narrowest chancel arch in any 
church under notice, and can in this respect only be paralleled 
in certain oratories of primitive type though of uncertain 
date in Orkney and Caithness, referred to on a previous page 
(ante, p. 127). 

The jambs and archivolt both in this arch and in the door- 
ways exhibit the characteristic strip-work already illustrated. 
The treatment of the archivolt of the chancel arch reminds 
us ot the classical ionic architrave with its three fasciae. 
The reeded pilaster of the north door may be compared with 
the pilasters similarly treated on the exterior. All the imposts 
are square and unadorned. There are three windows preserving 
more or less their primitive form but only that in the chancel 
is really original. That in the south wall of the nave is a 
restoration and the porch window has been altered. These 
windows are all splayed outside and in, and the original 
window in the chancel has sloping jambs. 

The last feature to notice is the occurrence high up in 
the eastern wall of the nave above the chancel arch of two 
figures of angels sculptured in low relief, in a style that shows 
little plastic feeling and might almost be called incised work. 
They are hovering horizontally in the air each holding over 
the two arms a napkin. They are the most important 
or at any rate the best preserved examples of Saxon figure 


sculpture in its connection with architecture, and form no 
doubt a portion of a lost group, a figure of the Crucified 
originally forming the centre. As has been previously ex- 
plained, no attempt is made in this volume to deal with Anglo- 
Saxon sculpture except in the tectonic forms of carved capitals 
and similar details. Any discussion of the Bradford angels 
must therefore be reserved, but it is well to mention that 
Canon Jones stated that they were ' found embedded in the 
wall ' above the chancel arch one on each side, and when he 
wrote in 1858 they were then ' placed over a wooden porch 
which has been erected as an entrance to the building on the 
west side.' ^ The reliefs are accordingly not now in situ but 
were placed where they are now at the restoration of the 

1 Wiltshire MagaxinCy loc. cit. p. 4.49. 




VII. The Western Tower. 

We come now to one of the most important features of 
this phase of architecture and we pass naturally to this from 
the consideration of the western porch, for it is one of the 
peculiarities of Saxon church architecture that in some cases 
a tower was built upon the walls of a previously existing 
porch. The most conspicuous instance occurs at Monk- 
wearmouth in county Durham. It will be remembered that 
about 675-80 A.D, Benedict Biscop founded two monasteries 
in the north, one at the mouth of the Wear and another 
not far away at Jarrow-upon-Tyne. Monkwearmouth, as it 
is now called, is joined to Sunderland by a bridge over the 
Wear and is a busy place of factories and shipyards. In 
the centre of it, and in surroundings which have not lost their 
old-fashioned aspect, there stands in an ample burial ground 
the church of St. Peter that presents one of the most curious 
illustrations of the abnormal character of Saxon work, and 
is a building over which a little time may profitably be 

The Saxon parts of the present church are shown in plan in 
Fig. 78. The nave, it will be noted, is of abnormal length 



measuring about 65 ft. by a width of nearly 19 ft. It is 
true that the Saxon chancel and chancel arch are eone, but 
the southern jamb of the present chancel arch has Early Norman 
character, and it is impossible to believe that the Norman 
builders who had no penchant for elongated plans had actually 
lengthened the already, to them, abnormal proportions of 
the Saxon nave. The present chancel arch must be in the 
same place as the Saxon one or if there have been any alteration 
it has been brought further westward. The west wall of 
the nave is original and is a little over 2 ft. in thickness. The 
south wall has been rebuilt on the original lines. The north 

Fig. 78. — Plan of the Church of St. Peter, Monkwearmouth, Durham. 

wall has been replaced by an arcade giving access to an aisle, 
but the western quoin of it is still visible on the exterior. 
The technique is rubble-work of rudely squared stones and at 
the corners the stones are larger and more carefully squared 
and fitted but there is no long-and-short or big-stone quoining. 
Height is again here a feature, for the west wall of the church 
(Fig. 79) measures externally 23 ft. 6 in. in width, and 31 ft. 
in height to the beginning of the slope of the gable, which ran 
upwards at an acute angle with the horizontal of about fifty 

The drawing gives the line of the original gable on the 
north side as it was seen in 1865, before the present north 
aisle was built. In the centre, an arched doorway 3 ft. 6 in, 
in width gives admission to the interior, and to the west of 

Fig. 79. — Western tower of Church of St. Peter, Monkwearmouth, Durham, 

before the last restoration. 



this is built a porch, measuring internally 8 ft. from north to 
south by 9 ft, 5 in. from east to west, and covered with a 
barrel vault of stonework running east and west, 12 ft. 6 in. 
high to the crown. There are doors to the porch on all four 
faces. Those to the north and south, 2 ft. 6 in. in width, 

FiG. 80. — Western doorway of porch, Monkwearmouth. 

have jtheir jambs rebated and heads splayed for doors opening 
outwards. The east door, also with a rebate,^ leads into the 
church. The western archway, 4 ft. 10 in. in width, which 
has jnever been closed by a door, is of extremely elaborate 
construction, and gives its stamp to the whole work (Fig. 80). 

^The rebate reduces the aperture of the doors on the north, south, and 
east of the porch to somewhat smaller dimensions than those given. 



Its jambs are composed of upright slabs lining the opening, 
surmounted by other slabs laid flat and bonding into 
the wall. On the surfaces thus formed there is carved 
on each side of the doorway an ornament consisting of a 

pair of serpents inter- 
twined, after a fashion 
which will be understood 
from the drawing, Fig. 
8 1, which shows the 
condition of the slabs 
when first uncovered in 
1865. On these slabs, 
as on a plinth, stand on 
each side two stone 
shafts, about 21 in. high 
by 10 in. in diameter, 
ornamented with an ela- 
borate system of projec- 
tions and cuts, that have 
evidently been produced 
in a lathe. These turned 
mouldings are far superior 
in their delicacy and accu- 
rate cutting (Fig. 82) to 
the general run of Saxon 
baluster shafts. 

These twin colonnettes 

Fig. 81. — Southern jamb of doorway to 
porch, Monkwearmouth, Durham. 

carry a massive impost 
1 1 in. high, chamfered beneath, and worked on all its edges with 
a roll moulding. From these springs the arch, formed of nine 
carefully cut voussoir stones of varying sizes, running right 
through the thickness of the arch, and once recessed on both 
outer and inner face, after the manner of the ionic architrave. 
The arris between the face and the soffit of the arch is 
worked into a roll like that on the edges of the imposts. 



Above the arch, at a height of 13 ft. 3 in. above the ground, 
the face of the porch is enriched with a flat string course, 
composed of various panels framed with cable mouldings, 
and carved in low relief with representations of four-footed 
animals, and at least one human figure. 

At a higher level comes a com- 
paratively large window, giving 
light to a chamber over the porch. 
The date of this window is problem- 
atical ; it is slightly splayed inter- 
nally, from an aperture of 2 ft. 6 in. 
to an inner width of 2 ft. 10 in., 
and has a cable moulding worked 
on the inside arris as in other parts 
of the porch. The outer arch has 
in any case been modernized, but 
the inner one looks antique. From 
this chamber also a narrow door- 
way, now the only means of access 
to the interior of the tower, forms 
a communication with the nave of 
the church. The aperture of this 
doorway towards the nave, only i ft. 
5 in. wide, has been modernized, 
but within the chamber where it 
is 2 ft. 6 in. wide its head was cut 

in a single big stone, and the construction of it shows 
clearly that it is an original feature. On the northern 
external face of the tower can be seen the marks of a square- 
headed doorway now blocked, that is at a height which would 
admit of access from the exterior to the chamber above 
the vault. This doorway however does not seem part of the 
original work. In connection with this a view of the church 
in the Gentleman's Magazine^ vol. Lxxxii/2, p. 513, may be 


Fig. 82. — Turned shaft from 
jamb of western door to 
porch, Monkwearmouth, 


Returning to the western face of the porch on the exterior, 
we notice above the large window a second string course, 
and from this level, at a height of 21 ft. above the ground, 
began the slope of the original gable that surmounted the 
porch. Under the apex of this gable there are five large 
stones let into the wall, the uppermost of which projects 
like a semicircular disc set horizontally, while the second 
shows the outline and shape of a human head, the tip of 
the lobe of the left ear being still, on a close inspection, 
visible. The two large stones next below have been hacked 
away flush with the wall, and the lowest has evidently been 
renewed in more modern times. It is clear that there was a 
statue here in high relief, about 6 ft. in height, and as it is 
placed with reference to the original gable, and has about it 
no marks of being a later insertion, it is presumably original. 

At this point occurs the junction of the original work of the 
vaulted porch with chamber over it, and the later tower reared 
upon its walls. On the western side of the tower as it at 
present stands there are clearly seen the sloping lines of the 
original gable over the porch, and that the upper part of the 
tower is a later addition is further proved by the fact that, to a 
great extent, it blocks out the light from the two original 
windows in the western wall of the nave. These windows 
are seen in the sketch of the interior (Fig. 83). The tower 
walls outside rise in front of these windows, and light is only 
admitted by splaying away the edge of the tower wall where 
it abuts on that of the nave, leaving an upright slit, the 
position of which is marked by the arrow at A, in Fig. 79. It 
must be remarked about these windows that they are splayed 
in the interior only, the aperture for light being i ft. 8 in. 
across and the width of the internal splay 2 ft. 9 in., and 
they have the peculiarity that in the lower part of their jambs, 
underneath the upright slabs that form these jambs, there are 
set baluster shafts of a kind similar to those in the porch 
though not so elaborate in execution. They are built firmly 



into the corners of the jambs and have their bases on a narrow 
ledge from which the sloping sill of the window rises at their 
back. They are the same 
height as the slope of the 
sills, so that the top of them 
comes nearly on a level with 
the bottom of the actual 
aperture of the window. 

The existence of these 
shafts was only made known 
at the restoration of 1866,^ 
previous to which the win- 
dows were covered up with 
plaster. It must be assumed 
that they are in their original 
position, and were intended 
as counterparts to the 
balusters in the external 
opening of the porch. The 
effect of them in the window 
jambs is however not very 
satisfactory. The upper 
part of the tower possesses 
features common to the 
numerous square western 
towers in other parts which 
will presently be noticed. 

This is not properly the place for any discussion of the 
relative dates of the various portions of the structure, but 
so important is the question whether or not the porch and 

^ The Ecclesiologist (Cambridge Camden Society) for 1866, p. 362, and the 
Transactions of the Architectural and Archaeological Society of Durham and 
Northumberland, vol. i, p. 141, and appendix, contain notices of the church at 
the time of these restorations, with many valuable illustrations. Mr. Hodgcs's 
paper in The Reliquary for July 1893 gives a good account of the church. 

Fig. 83. — Sketch of the western wall of 
the nave, Monkwearmouth, from 
the interior. 


the nave walls are the original work of Benedict Biscop in 
the seventh century, or at any rate near his time, that a brief 
note on the history of the site may here be permitted. 

It has been already noticed that we possess in the writings 
of Bede an almost contemporary notice of the first build- 
ing of the abbey of Wearmouth,^ Bede's account has 
been so often quoted that it is sufficient to say here that 
Benedict Biscop, founder of the monastery, in the year 675 
A.D. went over to Gaul and brought back with him some 
masons to build for him a stone church ' after the manner 
of the Romans in which he ever took delight.' To Gaul 
moreover he sent for certain workers in glass ; and for 
sundry fittings of the church not elsewhere to be procured 
he went himself on more than one journey to Rome. The 
fabric of the church must have been slight, for it only 
took a year to build ; it was on the other hand tolerably 
spacious, for even in Bede's time the community of monks 
in the combined monasteries of Monkwearmouth and Jarrow, 
of which the former was the chief, numbered 600 souls. 
It was not the only church in the monastery however, 
for we read also of a church of Our Lady and an oratory 
of St. Laurence, but it formed the place of assembly of the 
monks on important occasions, and was profusely adorned 
with several sets of paintings, had glazed windows, and a 
full equipment of sacred vessels, fittings, and vestments. 

The question naturally arises whether any of the work 
of Benedict's masons actually survives. We know that the 
sister establishments at Wearmouth and Jarrow flourished 
until the time of the disastrous inroads of the Danes, 
who, in the year 867, plundered and almost destroyed 
the churches and monasteries of Northumbria. From that 
date to the time of the Norman Conquest they disappear 
from the pages of history, but shortly after the Conquest 
they are, for a time, revived, and become once more the 
"^ Historia Abbatum, c. 5, Plummer's Bede, i, 368. 


seats of monastic communities. In 1083 however, the 
monks from these restored monasteries are transferred to 
Durham, and the houses of Wearmouth and Jarrow sink 
to the position of unimportant subordinate cells to the 
great cathedral abbey. 

It is of course conceivable that the church of Benedict 
Biscop, the rapid erection of which has just been referred to, 
might require rebuilding in the two hundred years between 
its first erection and the Danish invasion. We may 
accordingly ascribe the work or portions of it (i) to the 
founder's own time, (2) to some epoch in the two succeeding 
centuries, (3) to the period between 867 and the Norman 
Conquest, or, again, (4) to the era of revival from about 1075 
to 1083. With this latest date would accord the style of 
the Early Norman pier of the present chancel arch at the 
east end of the building, the half-columns of which possess 
the curious bulbous base which occurs in other examples 
of eleventh-century work in England, such as the crypt at 
Lastingham and the slype at Worcester ; while, though the 
plan, with the three half-columns, appears Saxon, the masonry 
and tooling are characteristically Norman. The upper part 
of the tower, on the other hand, agrees closely in style with 
the numerous examples of square western towers, occurring 
in Northumberland, Lincolnshire, and other parts, which are 
not Norman In character, and are generally ascribed to the 
first half of the eleventh century. Its total height is 60 
ft., which, on a face width of a little over 1 1 ft., repre- 
sents proportions far taller and more slender than are 
normal in English Norman architecture. It is not reduced 
by sets-ofF, nor was it intended to taper, though the weight 
of the tower bearing on the western archway has caused 
it to spread slightly, and made the tower measure an inch 
or two more at the base than at the summit. This upper 
portion of the tower appears earlier than the chancel arch, but 
it is quite possible that the two are practically contemporary, 


and coaeval too with the tower at Jarrow, which is Norman 
rather than Saxon in character. 

The questions of date now narrow themselves down to 
the query whether the lower portion of the tower and west end 
of the nave, which we have seen to be prior to the upper part 
of the tower, are due to an earlier restoration after 867 but 
before the Conquest, or are relics of the pre-Danish period. 
Now, Simeon of Durham, writing in Northumbria in the 
early years of the twelfth century, describes the havoc 
wrought by the Danes as so extensive that the monas- 
teries of the province were reduced to a desert condition, 
only bare walls being left, and the very sites of some 
passing out of knowledge altogether. The revival of 1075 
was brought about by a pious pilgrimage undertaken by 
some monks from the south of England, who ' had learned 
from the history of the Angles how that the province of 
the Northumbrians was formerly the home of crowds of 
saintly monks,' and who wished to visit the holy but now 
deserted sites,^ When the country had been reduced to 
such a condition, it is almost certain that any rebuilding 
of the churches ruined by the Danes would have been 
of a somewhat perfunctory character. The earliest work 
at Monkwearmouth is however marked by extreme care 
and elaboration in detail. The baluster shafts, the inter- 
laced serpents, the roll mouldings, the cable mouldings, 
the carved frieze of animals, the big statue in the gable, 
the balusters in the window jambs, are not everyday 
work, but represent quite the most extensive collection 
of carefully wrought details to be found in the whole 
range of extant Saxon buildings. It would have been 
practically impossible in that region between 867 and the 
eleventh century, while for its ingenuity and thorough- 
ness it is exactly what we should expect either from the 
wealthy and enthusiastic Benedict, or from one or two of 
'^Hist. Dunelm. EccL, iii, 21, Rolls Series, No. 75/1, p. 108 f. 



his successors during the flourishing period of the foundation. 
It may be added that the style of the work, though in 
many respects puzzHng, agrees better with an early than with 
a later date, for the reason that none of the known later 
features, which have been connected above with Germany, 
make their appearance in it. On the other hand there is one 
feature that would suit the eighth or ninth better than the 
seventh century. This is the great height of the nave walls, 
which is a peculiarity found neither in the basilicas of Roman- 
ized lands nor in the Celtic oratories, but comes into vogue in 
parts of the Continent as well as in England in the times of 
unrest and danger which fell upon Christendom when the 
Vikings forced their keels up the rivers of Western and Central 
Europe. Lofty walls and small apertures high up in them were 
a means of protection against raiders. In this case of Monk- 
wearmouth perhaps the suggestion of a rebuilding of Benedict's 
original structure before 867 will best meet the probabilities of 

the situation. 

There is one other certain case of a western tower built 

over an earlier porch and one or more in which the case is 

doubtful, while there is one 

instance known from literary 

sources in which it seems likely 

that a tower was reared on the 

walls of an existing lateral porch. 

The certain case is Corbridge 

above the Tyne valley, near the 

point where one of the great 

Roman roads to the far north 

crossed the Wall on its course 

into Caledonia. The pre- 
Conquest church, the nave of 

which survives, was entered through an ample western 

porch, the plan of which, reproduced by permission from 

one by Mr. C. C. Hodges, is given in Fig. 84. The 

Fig. 84. — Plan of tower and 
western end of nave, Cor- 
bridge, Northumberland. 


following description is taken from that by Mr. Hodges in 
the Reliquary of January 1893. 'The porch was entered 
by a round-headed doorway, five feet wide, and more than 
nine feet high, having a semi-circular arch. Above this is 
another semi-circular arch which is partly constructional, as 
it serves as a relieving arch, and partly ornamental, as its 
voussoirs are ornamented with a very early example of chevron 
work. This consists of a row of saltires, one on each voussoir, 
which vary in size. Immediately over the doorway is a small 
window with round head, widely splayed on the inside. 
Between the porch and the nave is a great entrance archway, 
eight feet two inches wide, and sixteen feet high (Fig. 85). 
The jambs are quite plain, and are formed of enormous stones, 
each one of which is as long as the wall is thick, so that 
there are no vertical joints in the jambs. At a height of ten 
feet six inches from the floor are projecting impost stones 
ten inches thick. These are of different sections on the 
two sides, and are Roman mouldings re-used, having been 
taken from the base or cornice of some great building. The 
arch is stilted to the extent of a foot. The voussoirs, of 
which there are thirteen, above the springing line, are two 
feet four inches long, and go right through the wall ; they 
are, however, three inches thinner than the wall. This 
difference is left as a recess on the east side, but does not 
extend to the two stones immediately above the impost which 
stilt the arch. It is clear from this, as from the dressing of 
the stones, that the arch has been bodily transferred from a 
Roman gateway, and merely re-set in its present position. 
The surrounding walls are almost, if not entirely, of Roman 
worked stone. Cramp holes and grooves, lewis holes, and 
broached tooling are everywhere visible, and the wavy, uneven 
surface of the walls, now that they are denuded of their 
plaster, although built of large square stones, shows that 
these did not always fit the thickness of the wall they were 
being built into. 



' The nave was about forty-eight feet long, and seventeen 
feet eight inches wide, and about twenty-nine feet high to 

Fig. 85. — Western portal, Corbridge, Northumberland. 

the wall-head. It was lighted by three windows on either 
side, one of which remains entire, but a considerable portion 
of the heads of two of them are still in situ in the north wall. 


From these remains we gather that they were of the same 
form and dimension as the still perfect window over the 
west door. The heads are in two stones only, one forming 
the internal arch and the other the external arch. The 
jambs, however, are formed of long through-stones. The 
inner head stones are by far the larger, as the splay is 
considerable. ... 

' The ancient chancel, and the arch opening to it, have 
entirely disappeared, and the indications to show the termina- 
tions of the nave in this direction are but slight. The west 
gable, however, remains entire, and shows that the roof, which 
seems to have been of thatch, probably ling, was of very high 
pitch, its ridge being nearly fifty-two feet from the ground. 

'At some period before the Conquest the roof of the porch 
was taken off and the gable removed. The side walls were 
then carried sheer up beyond the gable of the nave to form 
a tower. The north, south, and west walls of this were built 
on the walls of the west porch, but the east wall rested on 
the western gable of the nave. It is fortunate that this 
gable was not removed like that of the porch, or we should 
never have known how the change had been effected. The 
modern roof of the nave is much lower, both at the eaves 
and the ridge, than that of the early church, and portions of 
the west wall of the old nave now flank the tower, like 
buttresses which seem to rise up through the roof. The 
old gable window is now above the roof, and what is now 
its external side was formerly inside, and beneath the old 
roof, and its original external face is now to be seen inside the 

To this description from the pen of Mr. Hodges there 
only needs to be added the note that in the case of the 
window in the west gable of the nave just mentioned we 
find presented what looks at first sight like a double splay. 
There is the wide external splay which was originally internal, 
and a smaller splay on what is now the inner face of the 


wall. This, if original, would make the window a double- 
splayed one and there are chronological reasons which would 
render this very improbable. An examination of the window 
shows however that this inner splay has merely been hacked 
away in more recent times to let in additional light to the 
ringing-chamber. The window is really a single- or internally- 
splayed one like those in the north wall of the nave. 

The western arch, Fig. 85, is one of the most remarkable 
features of the kind in England. Its great height, and the 
large stones of which it is composed, give it an imposing 
aspect that its absolute plainness only increases. All the 
stones, whether in jambs or arch, go through the whole 
thickness of the wall. The technical peculiarities noted in 
the description just given make it practically certain that 
the arch at any rate if not the jambs is Roman re-used. In 
connection with the doubtful question whether or not the 
chancel arch at Escomb is also Roman re-used, it must be 
observed that the Corbridge arch does not show in the jambs 
the slab-like stones set alternately upright and flat, which 
are characteristic of the Escomb arch with other Saxon 
examples, such as the doorways at Monkwearmouth. 

The proportions of the plan of the Corbridge porch-tower 
should be noticed. It measures internally from west to east 
in the mean 11 ft. 4 in. and from south to north only 10 ft. 
1 1 in. At Monkwearmouth the corresponding measurements 
are 9 ft. 5 in. and 8 ft. These proportions accord with those 
of unaltered Saxon porches in which the dimension along 
the axis of entrance is always greater than that at right 
angles to this. Thus Bishopstone porch measures 12 ft. 5 in. 
by 9 ft. 2 in., Bradford-on-Avon porch 10 ft. 6 in. by 10 ft. 
2 in. (both in the mean), St. Pancras, Canterbury, western 
porch 10 ft. 7 in. by 9 ft. 6 in. The two known examples 
of towers reared on porches make it worth while inquiring 
whether this may not have been done in other Saxon examples 
where the base of the existing tower measures more along the 


line of access than in the other direction. The most promising 
instance is Bardsey, near Leeds, where we have a western 
tower joined to a church which possesses several early- 
features. The walls of this tower on the ground level are 
barely i ft. thick and the space they enclose measures lo ft. 
2 in. from west to east by 8 ft. from south to north. There 
is a narrow north door and old round-headed lights above 
it to north and south that are internally splayed and have 

sloping jambs, Fig. 86. Exter- 
nally on the western face there 
are marks of a gable similar to 
though not so pronounced as 
those at Monkwearmouth,' while 
the quoins at the lower part of 
the tower, when compared with 

those higher up, are of different 

Fig. 86. — Plan of tower and j r l ^ a • ^ 

, „ , ^, , and earlier character. As agamst 

western end, Bardsey Church, . ,. . , , 

near Leeds these mdications that the lower 

stage of the tower was originally 
a porch must be set the fact that there is no indication 
of a western door. Brixworth has been claimed as another 
instance of this kind, though there is no indication of a gable 
or sign of a break in the work between the lowest and second 
stages of the western tower. The plan here shows dimensions 
that are greater from south to north than they are along 
the line of entrance, thus reversing the arrangement just 
noted as normal. See Fig. 151, postea, p. 248. 

On the other hand there are towers which present no 
appearance of having been porches but which yet have the 
porch form of plan. The western tower at Ledsham, 
between Selby and Leeds, Yorkshire, measures 12 ft. 3 in. 
west to east by 9 ft. 8 in. in the other direction. It has 
a south door but none on the west. Middleton by Pickering, 
Yorkshire, which has a western door, gives us west to east 1 1 ft. 
10 in. by 10 ft. 6 in., and Wharram-le-Street, by Malton in 


the same county, 11 ft. 5 in. by 10 ft. 6 in. with a western 
doorway. Deerhurst western tower is exceptional in plan. It 
measures nearly twice as much west to east as it does south 
to north, but it is divided into two by a cross wall running 
in the direction last given (Fig. 169, postea, p. 299). At 
Bosham, Sussex, a western tower without any external door- 
way to it, measures internally 19 ft. 2 in. west to east by a 
mean of 15 ft. 7 in. north to south. Hence it follows that 
internal proportions are no criteria as to whether a structure 
is ab origine porch or tower. 

The original Romano-British or Saxon cathedral at Canter- 
bury, the plan of which drawn by Professor Willis from literary 
sources is so well-established, that it is generally treated as 
a monument,^ possessed before the Conquest two towers on 
its flanks halfway along the nave. The one on the south 
offered the principal entrance to the church and was dedicated 
to St. Gregory. The probability is that this was originally a 
porch over which the tower was afterwards built.^ It would 
have been impossible for Romanized Britons or Saxon Chris- 
tians of the first generation to have planned these flanking 
towers which do not belong to the architectural ideas of 
their time, but lateral porches of entrance would be quite 
in accordance with early Saxon habits. 

Passing now from the porch-tower to the tower proper 
we reach one of the outstanding features of our early 
architecture. The general aspect of the pre-Conquest 
tower has already been indicated. It will as a rule rise 
plain and unbuttressed and possess belfry openings of the 
form described on page 6 2- Western towers answering 
to this general description exist in some abundance and 

^ This building will be treated of in the sequel, postea, p. 260 f. 

2 The fact that this part of the structure was called by the vernacular 
term ' Suthdure ' (south door), a word incorporated in the mediaeval Latin 
of the writer who tells us much that we know about the church, shows 
that it was envisaged as a porch not as a tower. See Willis, p. 10. 


are most frequently to be met with in the eastern parts 
of the country especially in Lincolnshire. Referring to 
the map at p. 344 the three eastern Districts numbered in, 
VII and VIII supply about sixty-five examples while the 
rest of England furnishes hardly a score. Of the Saxon 
churches which are dotted so thickly about Lincolnshire, 
a county which has more pre-Conquest work to show 
than any other, three fourths, or thirty out of about two- 
score examples, have western towers. So characteristic of 
Lincolnshire is the form that the name ' Lincolnshire bell- 
towers ' is sometimes given to the whole group. Such 
towers cannot however be regarded as, in a strict sense, 
peculiar to the eastern counties, for specimens with some or 
all of the characteristic marks are common further north, 
and occur sporadically in the other parts of the country, 
as at Wickham, Berks ; Bosham and Sompting, Sussex. 

Most of these western towers especially the Lincolnshire 
examples possess as their distinguishing feature the double 
belfry openings with mid-wall shafts, or where the belfry 
stage has been rebuilt the previous existence of such 
openings can be inferred from analogy. They are how- 
ever commonly wanting in the more usual pre-Conquest 
indications such as long-and-short quoins, pilaster strips, 
and double-splayed windows. Lest their pre-Conquest 
character should be doubted however, it may be pointed 
out that a few towers agreeing with the mass in their 
general aspect do show these special marks of origin, 
while the double or triple openings with the mid-wall 
shafts occur in a few cases in the nave walls of churches 
that can be pronounced for other reasons to be un- 
doubtedly Saxon. Brixworth ; St. Martin, Wareham, Dorset; 
Worth, Sussex ; Wing, Bucks, are examples. Accordingly 
the possession by a plain unbuttressed tower of this particular 
form of belfry opening gives it a Saxon character, though on 
the other hand it must be admitted that some of the Lincoln- 


shire examples, which exhibit this distinguishing mark, have 
other details that betoken a later date. Such examples are 
Boothby Pagnell and perhaps Harpswell, Lincolnshire. The 
truth is no doubt that this particular feature was so well 
established in this region before the Conquest that it continued 
in use after that epoch. There need be no hesitation however 
in accepting the feature as Saxon (see ante, p. 63). 

It will be convenient to deal in the first place with such 
towers as possess none of the special Saxon features except 
the belfry openings — these towers number about fifty — and to 
reserve for subsequent notice the few exceptional examples 
that do exhibit other pre-Conquest indications. 

A selection may be brought together to illustrate the form 
and details of these monuments. As the subject embraces 
many points of interest it may be advisable to draw up 
the following scheme of treatment. We deal then with 
(A) the general form, (B) the architectural treatment of 
the elevation, (C) the finish at the top of the towers, (D) 
the west door where it exists, (E) the smaller lights on 
the lower stages. Entering then the interior there will 
fall to be noticed (F) the tower arch, (G) the internal 
fittings including openings to the church above the tower 
arch, (H) the means of access to the upper stages, (I) 
the belfry and its openings, with the caps and shafts in 
the mid-wall work. 

(A) In their general form the majority of these towers 
are of a tall and slender shape, and this has come to be 
accepted as typical. On the other hand most Norman 
towers are of comparatively broad proportions, and the 
Saxon type is often set against the Norman as the slender 
against the sturdy. It is well to note therefore that Norman 
examples both in England and in the Duchy are sometimes 
of elongated proportions. Between Caen and the sea there 
is a group of these, of which Lion-sur-mer, Luc-sur-mer, 
and Ver, are characteristic specimens, and these can be 



paralleled in our own 
country, as for example 
at Weaverthorpe on the 
Yorkshire Wolds. 

The example shown in 
Fig. 87 is St. Peter-at- 
Gowts, or St. Peter-at- 
the- Watercourses, in the 
main street of Lincoln 
below the hill. It is a well 
known monument and 
offers the characteristics 
^,.^^ of the group in an epi- 
^^ tome. These have already 
been indicated (ante, p. 
82). The tower is tall 
and narrow and there is 
no intentional batter in 
it. The walls are in the 
main vertical though as 
a fact they draw in 
a little in a curious 
fashion just under 
the string course, as 
can be seen in the 
drawing. The west- 
ern doorway and the 
lower openings of 
the tower are of 
a rather advanced 
character, and pos- 
sess, for example, 
projecting hood 
moulds. The double 
:i"*i^^^^^S'^^^iriiH^iE ■'^liSSM'*'! belfry openings how- 

FiG. Sj. — Western tower, St. Peter-at-Gowts, Lincoln. 

Fig. 88. — Tower of Clapham Cluirch, near Bedford. 

(To face p. i6o.) 

Fig. 89. — Arcading at base of Western Tower, Branston, Lincolnshire. 

(To face p. 161.) 


ever are of the type already referred to as giving to this group 
of structures their distinguishing character, while the tower 
arch and other features seen in the interior are of Saxon type. 

The simplicity and rugged strength of these towers make 
them at times very imposing, and this effect is perhaps 
most marked In the case of a midland example the tower 
of Clapham church near Bedford, Fig. 88. 

(B) A very limited amount of what may be termed 
architectural treatment is lavished on these austerely simple 
unbuttressed walls. A plinth Is sometimes present in one 
of the forms noticed on p. 85 but ex hypothesi the special 
group under consideration Is devoid of either pilaster 
strips or long-and-short quoins. Considering how commonly 
these details were employed In the later Saxon period when 
these towers must have been in building this Is not a little 
curious. The only architectural feature that breaks the 
absolute plainness of the walls is the horizontal string 
course. Only one of these ' Lincolnshire ' towers, Great 
Hale near Sleaford, has no string course or other break 
In its perfectly plain outline ^ but in every other case in 
this group one at least Is found, its most usual position 
being just beneath the belfry openings. A second Is less 
common. Above each string course the tower sometimes 
contracts but this is not universal. St. Andrew, Bywell, 
by the Tyne, and Monkwearmouth (Fig. 79, ante, p. 142) 
are examples to the contrary. 

Enrichment by means of arcading is not unknown. 
Tasburgh, Norfolk — if It be really Saxon — displays this in 
somewhat timid fashion, but the example par excellence is 
that of Branston near Lincoln, where arcading of somewhat 
advanced Romanesque character occurs on the lower stage 
of a tower that shows everywhere else only the normal 
features of the type. It Is shown in Fig. 89. 

^ Deerhurst ; and St. Michael, Oxford, are also quite plain, but they do 
not belong strictly to this group as they have distinct pre-Conquest detail. 


(C) With regard to the finish at the top of the towers 
our evidence is but scanty, and there is no reason to 
suppose that any one scheme was in universal use. Only 
one Saxon tower has preserved its ancient finish, and this 
is Sompting in Sussex where a form of cap common in 
Germany and called the German Helm is employed. The 
sketch, Fig. 90 (A), will explain the form better than a 
description. One of the other towers has a distinct indica- 
tion that this finish was once applied to it. This is St. 
Benet, Cambridge, which has Saxon features in the form 
of long-and-short quoins up to the very summit of the 
present structure, and we should at first sight conjecture 
that it had in Saxon days what it possesses now, a flat top 
like that of some Early Norman towers in the Duchy such 
as Lion-sur-mer to the north of Caen, If we examine 
the view of the upper part of it however in Fig. 90 
(B) the central example on the plate, we see that a 
pilaster strip starting high up from a corbel ascends in 
the centre of each face, and a comparison with Sompting 
indicates that this strip once ran up to the point of a 
gable, and that the finish was a German Helm. 

There are some indications at Corbridge that the tower 
had its eastern and western walls surmounted by gables 
carrying between them a roof of saddle shape. A pyra- 
midal cap of stone is the finish of many early towers, as 
of the Irish round towers and of that at Ver in Normandy, 
and a very interesting example survives in our own country 
in the square Norman tower of the oratory on Priest- 
holm, or Puffin Island, just at the south-eastern extremity 
of Anglesea, Fig. 90 (C). This is a form of vaulting, 
the stones being placed in horizontal layers but in encorbel- 
ment, and requires nice construction to which our Saxon 
builders, inferior in this respect to the Irish, were pro- 
bably not equal, so that it is doubtful if this should be 
included among possible terminations to Saxon towers. 



'WuhiKli.- M 










U -^ 














• i-t 


































(D) A western door of entrance occurs or has left its 
traces in about half the towers. These doors are of 
different types and exhibit some very characteristic Saxon 
details. The exceptional towers — for there must be 
observed this distinction between the fifty-odd towers that 


/ • 


Fig. 91. — Western doorway of tower, Clee, Lincolnshire. 

are Saxon only in form and in the character of their belfry 
openings and the few examples that have also long-and- 
short work and pilaster strips, — the exceptional towers have 
doorways in which the use of the pilaster strip at the 
sides and round the arch is common, but those of the 
' Lincolnshire ' type have sometimes developed Romanesque 
features such as angle shafts and recessed arches. The 
doorway at Branston may possibly be a later insertion, but 



Kirkdale^ and Kirk Hammerton, both in Yorkshire, are 

genuine Saxon work. The scheme of them was shown in 

Fig. 49 A and B, ante, p. 98. A more characteristic form 

of doorway is found in a few examples in the northern 

division of Lincohishire, and this feature is of interest as 

it cannot be exactly paralleled elsewhere. Fig. 91 shows a 

good example from Clee near Grimsby. There is a simple 

dignity about this 

massive portal which 

agrees with the general 

character of this type 

of tower. It should 

be explained that the 

imposts project 2 in. 

to 3 in. from the wall 

face and the outer 

order or hood mould 

is flush with them. 

The middle order is 

recessed 2 in. below 

this, and the inner, 

which corresponds to 

the wall face, is ^ in. 


Much more rarely 
than on the west do we find doorways on the lateral faces of 
Saxon towers. Porch-towers, and others of an exceptional 
kind to be afterwards noticed, possess these, but they are 
not found in the particular group now under review. 

(E) The smaller lights in the lower stages of the towers are 
ex hypothesi not of the double-splayed Saxon form but are 
usually internally-splayed loops with narrow outside apertures 
the head of which, after a fashion that is Roman and Irish and 

^ The doorway at Kirkdale is really the west door of the church, not a tower 
door, but it is useful for comparison. 

Fk;. 92. — Keyhole loop, west face of tower, 
Clee, Lincolnshire. Height, 2 ft, 8 in. 


Norman as well as Saxon, is cut out of a single stone. One 
particular form of aperture has received the name of ' keyhole ' 
from its form shown in Fig. 92. It occurs in the same 
district as the doorways last mentioned, but it is found also 
exceptionally in Oxfordshire, where the aperture in some mid- 
wall slabs in the centre of double-splayed windows in the 
tower of Langford church has the same outline. 

(F) On entering the tower our attention is directed first 
to the arch opening from its ground story into the body 
of the church. This is generally, but by no means always, 
of lofty proportions, and it seems to have been always in 
old times open and not closed by a door. There is an 
exceptional arrangement at Leathly, near Otley, Yorkshire, 
in a tower which lacks the distinguishing marks of the 
present type, the double belfry openings, and should only 
be included doubtfully in any pre-Conquest list. Here 
there is no tower arch but a doorway 3 ft. 4 in. wide, 
closed with an iron-bound door, the sill of which is 3 ft. 
from the floor of the church. There is an external western 
door to the tower but it is doubtful whether it is original. 
These points are of significance in connection with the 
question whether these early towers were used like the Irish 
round towers for purposes of defence. It may be said 
generally that there is not the smallest actual indication of 
such a use. Since half the towers have western doorways 
to the exterior on the ground level, and all but Leathly 
have open tower arches, an enemy could always gain the 
ground story of the tower and burn or smoke out those 
on the upper stages. 

One exceptional tower arch, that at Corbridge, has already 
received attention. Many fine simply turned arches devoid 
of enrichment occur in the ' Lincolnshire ' towers and look 
almost Roman in their unpretending dignity. Clee possesses 
a fine one measuring 16 ft. 8 in. in height by a width of 
6 ft. 9 in., which matches in character the western door 


already illustrated. St. Peter-at-Gowts exhibits a still more 
imposing portal of a height of at least 20 ft. to the crown. 
Many of the round towers of East Anglia, some of which 
seem of pre-Conquest date, are remarkable for their very 
lofty tower arches. 

Plain chamfered imposts are almost universal, but it may 
be noted that in one or two instances Roman worked stones 
are used for the imposts of these arches. Such is the case 
at Warden, by the Tyne, Northumberland, and likewise at 
Alkborough, Lincolnshire, where the jambs of the arch have 
also Roman moulded stones for a base. The method of 
construction with through-stones noticed ante, p. 96, is fairly 
common, but there are cases in which an arch is constructed 
with facing stones and rubble filling, and yet seems to 
belong to the pre-Norman epoch. The jambs have in many 
instances plinths of simple section, like D in Fig. 31, ante, 
p. 85, which exists at Clee. As a rule the imposts and 
plinths are not returned along the lateral walls, but instances 
of this occur. 

(G) Internal fittings comprise those arrangements which 
seem to show that the tower was used for a dwelling place, 
though not, as we have seen (vol. i, p. 334), necessarily 
for the priest. 

It has been already suggested (vol. i, p. 359) that the 
occupant was as a rule the ostiarius or sacristan, v/ho kept 
the doors, safeguarded the relics, and attended to the bells 
of the church, being bidden in some cases to ' ly over 
nyghtes therin.' A lodging on the first story of the tower 
would keep him in touch with the bells, and give him a 
place of vantage from which to command the altar with its 
treasures at the other end of the building^. In two existing 
Saxon towers presently to be noticed, Deerhurst and Bosham, 
there are small apertures in the eastern wall of the tower 
that were probably intended for the purpose of affording a 
view in this direction. 



One of the most cogent pieces of evidence of former 
habitation in these chambers of pre-Conquest towers occurs at 
Skipwith, Yorks, where in the eastern wall of the ringing 
chamber is a shallow recess 3 ft. high and 3 ft. 5 in. wide, the 
sill of it about 2 ft. from the present floor. Its depth is 6 in. 
The jambs are formed by round shafts each in a single stone 
with square abaci, quite of pre-Conquest type. It is to be 

Fig. 93. — Recess in chamber in the western tower at Skipwith, Yorlcshire. 
At A is the section of the lintel of the recess. 

noted that a small double-splayed window has been specially 
formed in the southern wall of the tower just where it would 
throw light on whatever was kept in, or used at, the recess. 
The absence of any parallel elsewhere makes it difficult to 
conjecture the purpose of the arrangement. It may have 
been a receptacle for a relief, such as a carved rood, or a 
panel like those now preserved in Chichester cathedral. 
See Fig. 93. In one of the upper stages of the tower 
at Deerhurst there are two aumbry-like recesses in the 
north and south walls that are also signs of earlier habita- 


tion. They measure about 2 ft, in height by a width 
of 18 in. 

The various openings, which in some ot these towers are 
pretty numerous, are important as evidence for the manner 
in which the structures were used in the olden time. The 
prominence assumed by the western forebuildings of German 
churches has already been noticed (ante, p. ^2 ^0 ^^^ some 
of our Saxon western towers, such as those of Deerhurst, 
Barnack, and Brixworth, seem to have possessed a similar 
importance in relation to the whole structures of which 
they formed the frontispiece. The arrangements in the 
great majority of the towers were much simpler, but in 
almost every case the adjunct seems to have been em- 
ployed for other purposes than merely for the accommodation 
of bells. 

There are at present stories in the towers formed at 
different levels by wooden floors, the beams of which rest 
sometimes on projecting corbels of mediaeval date, as at 
Bardsey, Yorks. Only in the case of one existing Saxon 
tower does the lowest stage possess a contemporary stone vault.^ 
This is at Monkwearmouth, but the vault here belongs to 
the earlier porch and not to the tower. In every other 
case the lowest stage of the tower, readily accessible through 
the western door or the open tower arch, is or was only 
roofed with wood. Access to this and to the stages above 
is gained by wooden ladders, some of which (not of Saxon 
date) are excellent specimens of rude but solid wood work. 
The first story is generally now the ringing chamber, and 
from this into the church there commonly opens a doorway 
which as at present placed is a somewhat puzzling detail. 

These upper doorways in Saxon towers are features so 
familiar as to be almost universal, but they are in no connec- 
tion with anything to be seen at present within the churches. 

^ With later mediaeval vaults inserted in Saxon towers, as has been the case 
at Barnack, we have of course no concern. 



They cannot be regarded as merely apertures to afford a 
view of the interior, for at Deerhurst and Bosham there 
exist by the side of the doorways small openings or squints 
which seem to have had this very object (Fig. 94). The 

purpose of the door- 
ways is clearly to 
serve as means of 
entrance and exit, 
and they might con- 
ceivably be designed 
to afford access by 
means of a ladder to 
the chamber in the 
tower. At Monk- 
wearmouth, where 
the porch below is 
vaulted, the only 
access to the cham- 
ber above, which 
was part of the 
original porch, is by 
a ladder up to this 
doorway shown in 
the sketch. Fig. 83, 
ante, p. 147. It is 

Fig. 94. — Western end of nave (eastern face of possible that some 

tower), Bosham, Sussex. From a sketch by ^^ ^^^ doorwayS in 

the late J. T. Irvine. •' 

the towers generally 

were used in the same manner, though this can only have 

been the case when they were at no great height above the 

ground. In Fig. 95 are given outlines to scale of the 

internal features at the western end of the naves of three 

characteristic examples, (A) Bosham (B) St. Peter-at-Gowts, 

Lincoln, and (C) Deerhurst. Bosham and Deerhurst, which 

show the small squints for inspection, have the corresponding 



doorways on lines respectively 18 ft. and 16 ft. above the 
floor, but in the Lincoln example where the tower arch is very 
lofty the height is 26 ft., and a doorway at that elevation 
can hardly have been used for access from the floor of the 

There are two other suggestions that have been made 
about these doorways. One is that they communicated 


-■^ ^^.. 


1 1 



ro' -D-- 



n •' 

Fig. 95. — Outlines of western ends of three Saxon Churches with western 
towers, showing openings in the eastern faces of the towers. 

with spaces between the ceilings and the outer roofs of the 
churches. Chambers between under and upper roofs are 
as we have seen features in some of the Irish stone 
churches of native origin (see Fig. 13, ante, p. 26) and 
occur also over chancels in Norman work in England, as 
at Darenth, Kent ; Tickencote, Rutland ; and Compton, 
Surrey. There is, as we shall see in connection with 
Barton-on-Humber, evidence of this arrangement in some 
Saxon examples where the side walls of the church were 


low, but in the case of large churches with lofty side walls 
the doorways in question would not be at a sufficient eleva- 
tion. Unfortunately there are very few examples where a 
church with a western tower has preserved the original 
windows in the north and south walls of its nave. 
Avebury, Wilts, possesses original Saxon nave windows in 
two stages, round-headed internally-splayed openings, with 
apertures 3 ft. high, on a line 9 ft. from the ground ; and 
internally-splayed circular lights in the upper part of the walls 
which were of great proportionate height. There is here no 
Saxon western tower, so we cannot judge from the evidence 
of tower openings whether the church had an intermediate 
floor, with an upper chamber lighted by the circular windows 
and approached by one of these doorways from the tower. 
There are many instances on the other hand in which it is 
clear that these enigmatical doorways did not open into such 
an upper chamber, and this explanation would probably only 
apply in a limited number of cases. 

The third suggestion is that they afforded access from 
the tower to wooden galleries erected against the western 
ends of churches. It appears that this has been the case at 
Deerhurst, where some moulded stones that project from the 
western wall of the nave at a level just below the doorway 
may have had some connection with a gallery, and on the 
whole this theory may be accepted as the explanation which 
probably covers the largest number of instances. 

There are some towers which possess similar doorways to 
those under question at a much higher level, and about these 
there can be little doubt that they opened into spaces between 
an upper and an under roof. At Bosham (A) Fig. 95, the 
height of the nave walls, about 29 ft., and the position of the; 
circular lights in the north wall,^ seem to point to this explana- 

■^ It is a question whether the north and south walls of the nave at 
Bosham are Saxon, or are due to rebuilding when the pointed arcades 
giving access to the later aisles were constructed. The circular lights are 


tion of the upper doorway. At Deerhurst too, if we assume 
that the line of the original external roof is approximately 
indicated by the weather-tabling on the east face of the tower, 
as shown by the upper dotted lines in the drawing (C) Fig. 95, 
we get the high doorway into a position suitable for the same 

For what purpose these upper chambers were used is 
another question, into which it would not be advisable to 
enter in this place,- Nor can we discuss here the exact 
intention of certain chambers in western towers that open to 
the church not by a simple doorway or a squint, but by an 
ornately treated double or triple aperture. Deerhurst has 
one of these (C), Fig. 95, and there is another at Brixworth, 
Fig. 153, postea, p. 252. At Deerhurst the chamber in question 
is on an intermediate story between the lower and the upper 
doorways, and has in it the two aumbry-like recesses already 
noticed. The opening towards the church exhibits features 
which will be discussed on a later page. At Brixworth, 
where there is no doorway above the tower arch, the chamber 
is much lower. These chambers were evidently used for 
more dignified purposes than the mere lodging of the church 
officer, and we are reminded of the chambers that accom- 
modated personages of note in the western forebuildings of 
certain German churches, such as the Minster at Aachen, 
Gernrode, and Eginhard's basilica at Seligenstadt.^ 

of a form known in Saxon architecture, and in spacing they are out of 
relation to the arcades below. Hence they may be accepted, with reserve, 
as original. Bosham church is noted postea, p. 327. 

1 The present roof, of the Perpendicular epoch, cuts as will be seen 
right across the doorway. The weather-tabling mentioned in the text, 
though it may belong to a later mediaeval roof, seems to give a suitable 
line for the Saxon one. 

2 There is an indication of the possible uses of the various parts of these 
towers in Mr. Micklethwaite's paper on Saxon churches in the Archaeo- 
logical Journal, liii. 

'^ Repertorium Jiir Kunstwissenschaft, xi, 399. 


We have not completed yet the study of these tower 
apertures, for they occur on other faces besides that turned 
towards the church, and sometimes indicate the presence of 
adjuncts built up against the towers. Thus at Deerhurst 
there is a doorway 25 ft. above the ground on the western 
face, that looks as if it had opened once on to the roof of 
some western adjunct to the tower. At Netheravon in Wilt- 
shire there is a western tower, late Saxon in general style 
but with some Norman features, that has distinct indications 

of the existence on the western, 
northern, and southern faces of 
former adjuncts, the purpose of 
which is problematical. The 
plan, Fig. 96, shows the attach- 
ment of these lateral walls which 
are now broken away. On the 
northern face, about 17 ft. above 

Fig. 96~Netheravon TowT ^he ground, there is an opening 

cut like a doorway, but only 4 ft. 
9 in. high, that may have given on to the roof of one of 
these subsidiary buildings.^ At Warblington, Hampshire, 
a square tower of rude workmanship now embedded in a 
beautiful church of later date, has doorways of this kind on 
the north, south, and west faces at a height of about 15 ft. 
from the ground. 

Such openings when on the plain faces of towers may 
always indicate the former existence of some adjunct, but 
the case is very different when the face of the tower below 
the opening is somewhat elaborately enriched. This is the 
case at Earls Barton tower, and will be discussed presently 
in connection with that most remarkable of Saxon monu- 

(H) The usual means of access to the upper stages of these 

^ There is a paper on Netheravon by C. E. Ponting, F.S.A., in the Wiltshire 
Magazine, 1901, from which some features of the plan have been derived. 


towers are as we have seen wooden ladders, there are however 
four examples of spiral staircases of stone enclosed in half or 
three-quarter round turrets built up against the western wall of 
square Saxon towers. The instances are Brigstock and Brix- 
worth, Northamptonshire ; with Hough-on-the-Hill by 
Grantham, and Broughton near Brigg in Lincolnshire. The 
general appearance of the turrets can be seen in Fig. 149, 
postea, p. 246. One of the towers of 
the Lincolnshire group, Great Hale, 
by Sleaford, is exceptional in that it 
possesses a narrow turret stair in the 
thickness of the wall at the north-eastern 
angle of the tower. The stairway is 
only I ft. 4 in. wide, and the construc- 
tion is of a very rude and tentative 

kind. The triangle of masonry in the ^'°- 97— Belfry Stage, 

, , Great Hale, Lincoln- 

corner was not quite large enough to , . 

'^ . shire. 

hold the cylinder for the stair, and the 

tower bulges a little on the eastern face to give it room. 

The plan at the belfry stage is shown in Fig. 97. 

(I) The belfry stage is the most important part of the whole 
tower, for here are exhibited the features which give this type 
of monument its pre-Conquest stamp. The general character 
of the double belfry opening with its through-stone and 
mid-wall shaft has been already made clear, and it only remains 
now to give some attention to the remarkable forms we meet 
with in the work, especially in the case of the caps. 

The mid-wall shafts are generally plain and straight-sided,^ 
without tapering or entasis, in section circular or octagonal, 
and sometimes oblong — that is, measuring more in the direction 
of the thickness of the wall, a peculiarity agreeing with a 
feature of some of the caps to be afterwards noticed. In most 
cases the shafts are provided with capitals, and less frequently 

^ At Barton-on-Humber and Glentworth, Lincolnshire, there arc enriched 
shafts (not balusters), but these are exceptional. 


with bases, with which feature time has dealt more hardly than 
with the more sheltered caps. The antiquity of many of the 
existing capitals is doubtful, and in what follows dependence 
has only been placed on those examples that bear a decided 
imprint of age. 

The capitals, a selection of which is shown in Fig. 99, 
p. 1 80, may be roughly grouped under the two headings, 
' Cubical' and 'Volute' caps, and appear to be genuine specimens 
of native English carving of the later Saxon period. There 
is every appearance that they were made for the places they 
occupy, and in a large number of instances they are only 
decorated on the outer face and part of the sides, so that we 
can imagine them being actually carved in situ. There is no 
order of historical succession to be made out in the forms and 
motives, for though some are primitive and others advanced, 
the ruder examples are sometimes to be found in towers that 
would be placed late in the group. In dealing with the caps 
however they may, for convenience sake, be ranged according 
to an assumed scheme of development, the most primitive 
being first considered. 

No. I, from St. Mary-le-Wigford, Lincoln, might seem 
of dubious antiquity did not the same form occur in the 
undoubtedly ancient archway, clearly at first a chancel arch, 
between the tower and church at Broughton, near Brigg, in 
the same county, Fig. 128, postea, p. 213. It represents a rude 
method of cutting down the square of the top of the cap 
to the octagon of its base, and in a modified form we find 
the same device at Boothby-Pagnell, Lincolnshire, where, 
however, the chamfers are hollow. Some caps at Marton, 
near where the old Roman road from Lincoln to Doncaster 
crosses the Trent, exhibit a somewhat helpless procedure 
by which the square of the abacus is cut down to the round 
of the top of the circular shaft by starting to slope the 
sides away like an inverted pyramid, and then rounding 
the corners off till the form becomes that of an inverted 


cone. This may be regarded as an uninstructed attempt to 
deal with a problem that is perfectly solved in the normal 
cubical cap, which is represented in examples of undoubted 
antiquity at Glee, near Grimsby (No. 11). This cubical 
cap, formed by the interpenetration of a hemisphere and 
a cube, is probably, as we have seen, an importation 
from Germany, where the form becomes common in the 
eleventh century. In its distinctness and decision of shape 
and perfect fulfilment of conditions the mediaeval cubical 
cap is as good as any tectonic form of the Greeks, and 
is the most successful independent invention of the kind 
that we owe to the middle ages. The development of 
the later subdivided or scalloped cap from the simple cubical 
type can be followed in the Lincolnshire belfries. 

At Rothwell, in the Caistor district, we find mitred cubical 
caps, shown in No. iii, but the tooling on the stone suggests 
that they may date from a restoration of half a century ago. 
Such mitring, due to a desire to accentuate the divisions 
of the mass, would be sure to occur as a stage towards its 
further partition in the subdivided cubical caps shown in 
Nos. IV and v. The type in which each face of the cap is 
bounded below by two semicircles instead of one occurs in 
ancient work at Branston, by Lincoln (No. iv). These caps 
are elegant and well-wrought examples. The further sub- 
division into three, by which we reach the shape of the 
familiar scalloped cap, is represented in the tower of the well- 
known pre-Conquest church of Bracebridge on the outskirts 
of Lincoln. 

Fig. 98 shows this church and tower from the south 
east. A more modern south aisle has been added to the 
Saxon nave, the long-and-short quoins of which are seen 
at its eastern end where the aisle wall joins it, and it may 
here be noted that in some cases the adjacent quoins of a 
nave will show long-and-short work while the quoins of the 
tower have no indication of such treatment. This is the 



case here ; at St. Peter-at-Gowts ; at Rothwell, Lincolnshire, 
and other examples in that county. It might be argued from 
this that the tower was added at a later period to a pre- 
existing church, but against this is the fact that the tower 
arch and the doorway which opens above it in the western wall 
of the nave presuppose in each case a tower. It is hardly 
credible that these features, which are on the whole extremely 
constant ones, have been in each case inserted in the western 
walls of originally towerless churches. When towers have 
in this way been added to earlier churches, as at Staple and 
Westwell in Kent, the older openings in the end wall betray 
the fact of the alteration. In the drawing of Bracebridge 
the square tower with its belfry openings appears at the 
western end. In the westernmost of these openings, not 
seen in the drawing, we find the cap given in Fig. 99, 
No. VI, in which the middle section of the front or western 
face of the cap is enriched in the somewhat bizarre fashion 
shown in the drawing. It is a good illustration of the 
curious mixture of normal with fanciful motives which we find 
in these caps, and which betrays the hand of ingenious, but 
half-taught, carvers at work on forms imported from abroad. 

The series just noticed represents a continuous evolution, 
in which a form produced originally by the exigencies of 
construction — or, to use a convenient term familiar in 
Germany, a tectonic form — is gradually modified by sub- 
division and by the defining and accentuating of parts. The 
two dimensions, the square above and the circle below, 
depend upon structure, and the artistic problem lay in the 
fitting transition from the one to the other. The transition 
is, as we have seen, worked out in the Lincolnshire belfries 
in different ways, though we are not to conclude that the 
simpler and ruder forms are earlier in actual date than those 
more artistically advanced. 

Hitherto we have dealt with caps of the cubical type. 
Those which are distinguished by the use of the volute 










^^^^-.•i if 




might seem at first sight more artificial, as the volute itself 
is obviously a borrowed form. The volute however, as 
used in these belfries, is only a means of decorating a shape 
arrived at in the process of construction. The construction 
in this case is not the same as in that of the normal cubical 
cap. In the latter, the diameter of the hemisphere that 
interpenetrates with the cube is equal to the diagonal (in 
plan) of the cube. In the caps now before us (see Nos. 
VII and viii) the lower part of the cube is worked into the 
form of a hemisphere of a diameter equal only to the side 
of the cube. The smaller hemisphere is carried round in its 
full circumference for about half the height of the cap, at 
which point the corners of the original cube are left projecting. 
These projections have then to be dealt with, and they are 
brought down to meet the hemisphere in various ways, of which 
that shown in No. viii is the most common. The cap from 
Bracebridge, south opening (No. vii), is notable for its 
originality. The shaft supporting it possesses moreover 
about the best developed base that occurs in the belfries, 
the profile of which is given in the drawing. In the Glent- 
worth example (No. viii), and many others of which specimens 
are given in the illustrations, the projecting corners of the 
cube are worked into volutes, while the central space on each 
face between the curls is left plain, as in No, viii, or treated 
with a drop like the so-called Tau of Early Norman caps, or 
some other ornamental motive. Some of these volute caps 
introduce us to a more elaborate decorative treatment. At 
Scartho, near Grimsby (No. ix), the lower part of the capital 
is surrounded with a ring of upright leaves turned over at 
the top after a fashion represented in the Early Norman 
crypt at Lastingham, as well as in the crypt of Ste. Trinite, 
at Caen, Normandy. The richest in ornamentation of all 
is the cap from St. Peter-at-Gowts, south opening (No. x), 
where there is considerable elegance of design and sharp and 
delicate cutting. The cap from the upper stage at Barton- 


on-Humber, south opening (No. xi), is, on the other hand, 
clumsy and unpleasing, though undoubtedly original in treat- 
ment. It passes off into the octagonal shaft which it crowns, 
without any neck moulding, though this is almost universal 
in other examples. 

The most interesting set of caps in any of the towers is . 
to be found at Great Hale, near Sleaford, where there are 
four, all different and all fanciful without being extra- 
vagant. The caps are only carved on the outer faces, and 
they have the peculiarity that the abacus measures rather 
more in the direction of the thickness of the wall than it 
does the other way. This gives them some slight approach 
to the corbel capital, branching out to take the width of 
the masonry, specimens of which from different regions were 
shown in Fig. 26, ante, p. 64. In shape the Great Hale 
examples are a sort of combination of the cubical with the 
volute form, and they are peculiar in the reeding with 
which the lower part is adorned (Nos. xii, xiii, xiv, xv). 
The last example No. xvi, from the northern face of 
Glentworth tower, shows the extreme limit to which is 
carried the principle of corbelling out the cap to correspond 
with the thickness of the wall. This cap measures 1 1 in. 
on its face, but the side extends to 16 in. by a curious 
tongue projecting at the back. This extension is in none of 
the examples under consideration carried far enough to make 
it possible to dispense with a through-stone, but the corbel 
cap that is cap and through-stone in one does occur, though 
rarely, in our English work (Fig. 26, iv and vi, ante, p. 64). 

In connection with the subject of moulded caps an additional 
word may be said upon the treatment of imposts in Saxon 
building. These imposts, on the piers of chancel and tower 
arches and in the form of through-stones in belfry openings, 
are common and at times characteristic features. The most 
usual form of the plain chamfer is equally well represented in 
Norman (Fig. 47, ante, p. 96) and even in Roman work (Fig. 

S'. Marj-, V'l. Ikacebridgc, by Lincoln. \1I. Do., do. \ III. Glentwortli, 
I,inco I-, XIV., XV. Great Hale, near Sleaford. XVI. Glcntworth. 

(To fiice p. i8o.) 

Fig. 99. — Cipii.ils on mid-wall shafts in Western towers of the 'Lincolnshire typfi. 

I. S.. Mary-k-Wigford, Lincol„. 11. Ckc, Lincolmhire. III. Rothwell, b, Ca.stor. IV. Branston, ncnr Lincoln. \-., VI. Bracebridgc, by Lincoln. V" Do., 
Lincolnshire. IX. Scareho, near Grimsby. X. St. Peter-a,-Gow,s, Lincoln. XI. Barton^cn-Humbcr. XII., XIII., XIV., XV. Great Hale, near Sleaford. 

Jo. \ III. Glentworth, 
XVI. Glentworth. 

(To r4ci p. 180.) 



I, ante, p. 4), The hollow and the quirked chamfer introduce 
varieties. The former does occur as an unimpeachable pre- 
Conquest feature in the window in the east wall of the nave 
at Wing, Bucks, and appears in the chancel at Deerhurst and 
the belfry at St. Peter-at-Gowts, Lincoln. It occurs, too, on 

Fig. 100. — (Scale about -^ of nature.) 

A, A'. Barholm, Lincolnshire, jamb of south door of nave. 

B. Barnack, Northants, tower arch. (See Fig. 122, postea, p. 206.) 

C. Coin Rogers, Gloucestershire, chancel arch. 

D. Howe, Norfolk, tower arch. 

E. Deerhurst Chapel, chancel arch. 

F. Repton, Derbyshire, pier of vault ; F', wall-cornice. 

G. Pattishall. Northants, chancel arch. 

H. Corhampton, Hants, north door of nave. 
J. Daglingworth, Gloucestershire, south door of nave. 

one of the string courses at Earls Barton tower. Of the 
quirked chamfer the writer knows of no instance in work that 
is certainly Saxon, and the examples of it in towers of the 
' Lincolnshire ' type may be due to a Norman chisel. On 
the other hand, imposts in the form of huge trapezoidal blocks, 
as at Wittering, Northants (Fig. 59, ante, p. 108), and Market 
Overton, Rutland (Fig. 46, ante, p. 96), are unmistakably 
Saxon, as likewise those apparently formed of superimposed slabs 


(Barnacle, Northants, Fig. loo, b, and postea, p. 206 ; Miserden, 
Gloucestershire, Fig. 172, postea, p. 327). Moulded imposts 
must be judged by the character of the work and its surround- 
ings. A few specimens are given in Fig. 100. That at 
Corhampton (h) from the blocked north door of the nave is, 
on a small scale, like the grand impost of the tower arch at St. 
Benet, Cambridge. Deerhurst Chapel (e) exhibits the timid 
Saxon grooving. The other examples are more advanced, but 
even that at Coin Rogers, Gloucestershire (c), is in Saxon 
surroundings, and the arch and jambs are constructed with 

The most elaborate example of impost ornamentation occurs 
in the narrow doorway in the south wall of the nave at Barholm, 
Lincolnshire (a). The Saxon character of the work has been 
questioned, but the pre-Conquest date of the doorway is attested 
by the fact that above it there rises a portion of a Saxon pilaster 
strip, occupying the same position, just above the crown of the 
arch, that pilaster strips occupy in the undoubtedly Saxon 
examples of Corhampton, Hants, and Stanton Lacy, Shropshire, 
(See ante, p. 89.) The ornamentation on the top of the jambs 
and on the imposts is timidly incised, and agrees with a Late 
Saxon rather than a Norman date. The base moulding, which 
runs along the south wall of the nave, is of rather more orthodox 
character. The profile of it is given at a'. 

The towers of the ' Lincolnshire ' type hitherto dealt 
with are all of the square form. In East Anglia and 
Essex the characteristic plan is circular, the tower being 
embedded so to say for some part of its circumference in 
the western wall of the church but showing its completely 
round shape above. The significance of the form has been 
a good deal discussed,^ but there is really no mystery about 

1 There are papers on the Round Towers of this region in Archaeologla, 
XXIII ; Journal of Archaeological Association, xxi, xxxvii, xriv, xlvi, xlviii, ; and 
on Saxon Architecture in Norfolk in the Archaeological Journal, vi. 


it. The round form of church tower though characteristic 
of East Anglia and Essex is not, even as regards this country, 
pecuHar to that region, while the form occurs often enough in 
other lands. The round tower attached to the basilica of S. 
Apollinare-in-Classe by Ravenna is probably the earliest existing 
round tower,^ perhaps the earliest of all extant towers of any 
form, and the feature occurs elsewhere among the Ravenna 
churches. There are round towers on the Plan of St. Gall 
of about 820 A.D. and the form is frequent in Ireland from 
about the tenth century onwards, and overlaps into Scot- 
land. In England there are round towers associated with 
Norman detail, as at Haddiscoe, Norfolk, and Little Saxham, 
Suffolk ; of Early English date, at St. Michael, Lewes, and 
Hythe, Kent ; and in the Decorated style at West 
Shefford and Welford, Berks. It seems best to take the 
simplest explanation of the form that offers itself, and to 
regard it as due to the want of freestone suitable for 
quoins or to the absence of time or skill for squaring it. 
The Irish round towers probably owe their form to the 
conditions of haste and pressure connected with the Viking 
descents, while the English round towers are generally in 
regions where the building material is largely flint, and good 
workable stone is rare. This is notably the case in East 
Anglia and Essex where accordingly these round towers are 
at home. 

The absence of cut-stone details renders a decision as to 
the comparative date of these towers not always easy, but 
some of those in District III are almost certainly Saxon. 
Bessingham near Cromer, Witton by North Walsham, Great 
Ryburgh, East Lexham near Swaffham, Gissing by Diss, St. 
Julian at Norwich, Howe and Colney in the vicinity of the 
county town, all in Norfolk ; with Herringfleet, Suffolk, 

^The basilica was dedicated in 549 a.d. but opinions differ as to whether 
the campanile is contemporary with the rest of the fabric. There seems no 
strong reason to the contrary. 


have all pre-Conquest claims, but it must be noted that the 
double-splayed window which some of them present as their 
chief warrant is not so distinct a Saxon sign in East Anglia 
as it is in other parts of the country.^ Bessingham and 
Herringfleet are the best attested, as they possess besides 
mid-wall shafts the peculiar Saxon strip-work as a framing 
to some of their belfry openings. The plans Figs, loi, 
1 02 show the manner in which these towers are joined 
to their (originally) aisleless churches. 

Fig. 1 01. — Plan of Witton Church, Norfolk. 

The above are the principal points of interest about the 
particular class of monuments under notice, that is to say, 
the fifty odd towers of the ' Lincolnshire ' type that have 
only their general form and their belfry openings of Saxon 
character. The exceptional towers to which reference has 
already been made exhibit, over and above the double belfry 
openings, such features as long-and-short quoins, pilaster 
strips, and double-splayed windows. The relation of some 
of them to the plan of the church to which they belong 
presents also some remarkable features. 

The tower of Earls Barton church, Northamptonshire, 
figured on the frontispiece, is by far the most noteworthy 
architectural monument of the Saxon period. It is one of 
the few in which we can take an aesthetic as well as an 
antiquarian interest, for it represents an idea of some 

^ See the note on the index list of Saxon churches at the end of this volume. 


grandeur carried out with great pains and elaboration. It 
is a useful object lesson in the artistic capabilities and 
shortcomings of the Saxon builder. It has undeniable 
greatness not only of actual size — the tower measures 68 
ft. 8 in. to the top of the modern battlements by a width 
on the western face of a little over 24 ft. — but also of 
dignity of statement. The designer has made the best of 
the means at his disposal, and has employed most of the 
details available at his time in more exuberant fashion than 
in any other example about the country save perhaps the other 
great Northamptonshire tower of Barnack. Yet the effect 

Fig. 102. — Plan of Bessingham Church, Norfolk. 

is not truly architectural. The mass of the structure is fine 
but the details do not grow naturally out of the construc- 
tion, nor on the other hand are they in accordance with 
the grammatical employment of such details in architecture 
generally. Care has been taken in small points such as the 
finish of the upright pilaster strips under the first string 
course, but the base of the tower was set out so carelessly 
that the southern side is about a foot longer than the 
northern, while the south west angle is decidedly acute 
and the north west correspondingly obtuse, see plan Fig. 
103. Being however what it is, the most characteristic 
piece of Saxon work in the land, it is worth while 
examining it in some detail. 

From a simple square plinth rise the walls of rubble work 
plastered. They have a thickness on the ground story of 

1 86 


about 4 ft. but they decrease in thickness by the sets-off as they 
ascend, till at the belfry stage just under the battlements they 
measure 2 ft. 6. in. The faces are enriched by vertical and 
horizontal members and the corners are strongly accented. 

The vertical members are pilaster 
strips about 4 in. in width and 
they are joined at different 
heights by round or straight- 
sided arches arranged in a 
fanciful and illogical fashion. 
The round arches above the first 
string course have no sense in 
that position. The horizontal 
divisions are three string courses 
of which the first has a hollow 
chamfer, the other two being 
square in section. The quoins 
Fig. 103.— Plan of western tower, show pronounced long-and-short 
Earls Barton, Northamptonshire, work, and it may be noticed that 

the ' short ' pieces, or as they 
should rather be called the flat slabs, are not, as is often 
the case, cut back in a line with the ' long ' or upright 
pieces so that the plaster would up to this line conceal them. 
The present plastering of the tower is modern but when 
the original plastering was complete these pieces would 
always have been apparent. It should be remarked that the 
eastern quoins of the tower are as marked as the western and 
come right down to the ground, though the more modern 
nave has been built up against them. This point, it will be 
seen, is of some importance. 

The openings on the ground story are a tower arch that 
has been altered in Norman and in later times, and a 
characteristic western door of which Fig. 104 gives a view. 
The height to the crown of the arch is 8 ft. 7 in., the width 
between the jambs 3 ft. 3 in. The head of the doorway is on 



the exterior cut out of two stones, but in the interior the 
whole head is formed in a single huge block. The jambs are 


Fig. 104. — West doorway, Earls Barton Tower, Northants. 

formed by large slabs set upright alternating with flat ones 
after a fashion we have come to know. On the north side the 


slab which forms almost the whole height of the jamb is 
4 ft. 6 in. high, 6 in. thick and 3 ft. 7 in. in depth. The 
enrichment consists in an outer order of upright pilaster strips 
square in section that are bent round above in the shape of the 
arch, and two inner mouldings half-round in section. The 
imposts and the plinths are plain square blocks, but upon the 
face and sides of the former there is an incised enrichment 
consisting in a plain sort of arcading only cut in to a depth of 
about f^th of an inch. This arcading Mr. J. T. Irvine con- 
sidered to be an afterthought of a Norman carver. 

Above this doorway is an internally-splayed window of a 
common form that appears to be an insertion of Norman 
times, for it interferes with an older double window of the 
original work which is now blocked on the exterior, but which 
resembled the double opening that still exists on the southern 
face of the tower and is shown in the view (see frontispiece). 
These openings are double-splayed and the apertures on the 
western side were cut in mid-wall slabs in the form of a circle, 
while the apertures on the south are in the form of a cross with 
equal arms. It is worth noting that at East Lexham in Norfolk 
there is a mid-wall slab in a pre-Conquest opening in the tower 
that is also cruciform, only there the cross is stone and the back- 
ground aperture, while at Earls Barton it is the aperture that 
is cruciform. Above each of these southern openings is an 
enrichment of narrow roll mouldings disposed about a central 
cross carved in relief. 

On the intermediate stage where is the clock-face occur on 
every side but the north those enigmatical doorways apparently 
leading no whither, to which attention has been called. That 
on the eastern face of the edifice now gives access from the 
tower on to the comparatively flat roof of the nave. The roof 
of the Saxon nave was however of much higher pitch, and it 
has left its mark on the eastern face of the tower. The apex 
of the gable of the external roof reached a point corresponding 
with the middle of the height of the triangular headed opening 


seen in the view above the clock-face. Hence the round- 
headed doorway on the eastern face of the tower opened at 
one time under the external root of the church, and either com- 
manded a view of the interior, or else communicated with a 
chamber between the external and the internal roof of the nave. 
(See ante, p. 169 f.) In any case the height of it above the 
floor of the church, about 25 ft., must have been far too 
great to admit of that access by means of a ladder which 
has been suggested as one way in which these doorways 
may have been used. The doorway on the southern 
face, measuring 7 ft. by 2 ft. 6 in., offers egress from near 
the floor of the ringing chamber, but the external aperture 
is at a giddy height above the ground, and there is no apparent 
sign of any gallery or platform to which it may once have given 
access. The treatment of the face of the tower below the 
opening seems to preclude the idea that any adjunct was ever 
built up against it. The interpretation of these external door- 
ways which has been suggested in the cases of Deerhurst and 
Netheravon, ante, p. i 74, does not here apply, and the purpose 
of the apertures is a puzzle. The triangular-headed openings on 
the stage above are curious. Like the round-headed doorways, 
they are cut straight through the thickness of the wall with- 
out any splay. This is usual in the case of doorways but 
quite abnormal in that of window openings, which especially 
in Late Saxon work are always deeply splayed either internally 
or on both faces of the wall. 

We now come to the uppermost stage where the original 
work ends. Here is the bell-chamber, and for the free trans- 
mission of the sound there is on each face a group of five 
openings that are arranged in a fashion that gives us an 
interesting modification of the usual mid-wall work. The 
plan of the southern group of openings (Fig. 105) shows 
that the main part of the thickness of the wall is carried 
by simple square stone pillars, while the shafts that are intended 
to be seen are thrust forward to the external edge of the 



opening, the result being an arrangement resembling in a rude 
form the duplicated shafts in Early Christian buildings spoken of 
ante, p. 62. The same arrangement may be observed below in 
the case of the double-splayed lights with cruciform aper- 
tures. Here shafts of the same kind as those in the belfry 
openings above are pushed out on projecting corbels and 

employed as mere 
ornaments not sup- 
porting anything. 
This is only another 
instance of the il- 
logical treatment of 

features to which 
Fig. IOC. — Plan of part of belfry stage, Earls . , , , 

Barton. attention has already 

been called. 

The form of these shafts introduces us to a new and char- 
acteristic feature of Saxon architecture. They are what are 
termed ' baluster shafts,' that is short columns banded at 
intervals and swelling out between the encircling rings. The 
shafts in the belfry here are abnormal for they are mostly 
oblong in plan instead of round, and in this case have the 
mouldings cut only on the external half. As a rule such 
shafts are round, and this is indeed part of their very nature 
for the earliest and most elaborate of them were turned in 
a lathe. Earls Barton tower having again brought us into 
contact with the feature we have been introduced to at 
Monkwearmouth (ante, p. 144, 5) we will in accordance 
with the plan here adopted follow it throughout the districts 
wherever it appears. 

In dealing with the porch at Monkwearmouth we have 
had before us examples of the turned baluster shaft that are 
far more refined and delicate in workmanship than these at 
Earls Barton, and the relation of the two needs some ex- 
planation. As the whole subject is an interesting one it 
may be treated here in a connected manner, even though 



this will involve a certain amount of historical discussion 
which is in strictness beyond the scope of the present 

The mid-wall shafts dividing the belfry openings in church 
towers of the ' Lincolnshire ' type are usually quite plain. 
The true Saxon baluster shafts do not as a rule occur in these 
belfry openings, but they are met with in some of the ex- 
ceptional towers, as here at Earls Barton, or at Barton-on- 


Fig. 106. — Part of Roman altar from Birrens, Dumfriesshire, showing baluster 

shafts and astragal. 

Humber, and St. Benet, Cambridge. The accompanying 
drawings show some characteristic examples of these curious 
features of our pre-Conquest buildings, with which are 
connected some rather puzzling questions of origin. 

We may start with the hypothesis that they are early 
features, partly because they occur in the western porch at 
Monkwearmouth, and partly because a comparison with 
Roman remains indicates an early period for their appear- 
ance. In technique and in a certain general character they 
resemble similar features in Roman work, though on the 



Fig. 107. — Astragal on Roman stone 
at Hexham. 

Other hand they are in a most important respect entirely 
unclassical. The Roman evidence comes mainly from the 
North, and appears in the form of representations of baluster 
shafts on a small scale on sculptured stones, such as altars. 
For example, on a Roman altar found not long ago at Birrens, 
Dumfriesshire, on the Scottish side of the great Wall, we 
find the detail shown in Fig. 106, where in the centre of 
the front a round-headed niche is flanked by two supports 

that evidently represent such 

balusters. A small Roman 
altar from Lanchester, in the 
cathedral library at Durham, 
exhibits a niche flanked in the 
same manner. These upright 
shafts, represented as used constructively, must be distinguished 
from similar motives, strung together in a sort of beading, 
and used to form continuous lines of enrichment, as on the 
Birrens altar, on each side of the niche in Fig. 106. It is 
probably incorrect to speak of these as ' rows of balusters.' 
They are really forms of the 
astragal ornament, though 
the elements of the pattern 
may at times have been 
influenced in form by 
baluster shafts. In a Roman 
stone built into the north 
wall of the north passage 

of Wilfrid's crypt at Hexham, there is an astragal of a 
simple type (Fig. 107), and the Birrens beading would be 
much the same, only that in the middle, at the thickest part, 
each bead is either cut in two or marked with a nick_, it is 
not easy to say which. Now these same forms of supporting 
baluster and of beading occur in undoubtedly Saxon stones, 
many characteristic examples of which are in the Durham 
library collection. Fig. 108 shows one of several fragments 

Fig. 108. — Carved stone, with balusters, 
from the site of Wilfrid's Church 
at Hexham. 



found on the site of the nave of Wilfrid's church at Hexham, 
and there is some reason to believe that these formed part of 
its enrichment. Here distinct balusters occur as ornamental 
motives, while there are other stones in the same collection 
with Saxon carving on them, which show forms of the 
beading above noticed. On a well-known stone in the 
porch of Jarrow church (Fig, 109) there is a row of little 
balusters, about 3^ in. high, set upright, but curiously- 
similar in shape, if we look 
at them in one way, to the 
elements of the horizontal 
beading on the Birrens 

The Roman carver would 
probably not have repre- 
sented these baluster shafts 
as used constructively unless 
such features had been 
employed in real life, 
though nothing exactly like 
them may have been found 
on Roman sites. In Saxon work they are not only 
represented in ornament but actually occur used construc- 
tively in buildings. That they are of Roman derivation 
may be inferred from the comparisons just made between 
Roman and Saxon stones, and is rendered still more 
probable by the technical peculiarity that the Early Saxon 
balusters, like the small Roman shafts with moulded caps 
and bases, are turned in the lathe. We come now 
however to the striking fact that the oldest Saxon shafts 
occurring at Monkwearmouth, and in the work of coaeval date 
at Jarrow, are in profile quite unlike anything we find in 
classical architecture. The existing Roman shafts of small 
size found in this country (save only some of a special class 
to be noticed presently) are classical, insomuch as they exhibit 


Fig. 109. — Carved stone in porch of 
Jarrov/ Church. 


the three parts of the normal column — base, shaft, and capital. 
The base and capital may be worked into any number of 
mouldings, and the neck may be similarly treated, but the 
main divisions of the whole are not really obscured. The 
character of these Roman shafts is well illustrated by the ex- 
ample shown in Fig. 4, ante, p. 10 ; one germane to the 
present subject, as it actually occurs in the belfry opening 
of a Saxon church tower at Wickham, Berks. The church 
lies close to the Roman road that strikes off from the 
Kennet valley at Speen (the ancient Spinae) in the direction 
of Cirencester, and this may explain the appearance in the 
building of two unmistakably Roman shafts, of which one 
is given in the illustration. Here we have the normal 
elements — base, shaft, necking, and capital — of conventional 
late classical form. 

In the same way the decorative balusters already noticed, 
whether Roman or Saxon, have base, shaft, and capital. On 
the other hand, if we take actual Saxon balusters in real life, 
as illustrated in the accompanying drawings, this classical 
membering of the shaft is not common, and when it does 
occur it is in examples that may be placed comparatively late. 
The early shafts at Monkwearmouth and Jarrow are con- 
spicuous for the complete absence from them of any sign of 
this classical norm. They are decorated with a good deal of 
elaboration by means of numerous shallow projections and 
hollows that are distributed symmetrically above and below the 
centre, and take no account of head and shaft and base. The 
most important examples, those actually in situ in the jambs of 
the porch at Monkwearmouth, are illustrated in Fig. 82 
(ante, p. 145) ; but there exist there, or at Jarrow and 
Durham cathedral library, whole or in fragments, nearly 
fifty similar shafts of the same character, but varying in 
the distribution of the projections and hollows. Tlie score 
or more of them in Jarrow porch are shown in the upper 
and undermost row in the process plate Fig. no, while the 


(To face p. 194.) 

Fig. I 10.-— Baluster shafts at jarrow and Monkwcarmouth. 
(I'loni [)hotographs kindly fuinislu-d by J. I'aitison (iihson, IIl-xIi.iiu.) 


centre row shows some that are preserved in the vestry at 
Monkwearmouth. The Jarrow examples are 2 ft. 5 in. in 
height by about i ft. diameter, and they seem to belong to 
two types, though the examples of each type vary slightly. 
In one type the general outline is straight, while in the 
other, parts of the baluster swell out beyond the general line. 
This difference occurs also in the Monkwearmouth examples, 
which are less easy to group and rather smaller in size. In 
no case however, in this particular set of shafts, do we find 
the marked bellying out of the whole shaft towards the centre 
part, or the drawing in at the neck and base, of which the 
other illustrations which follow offer so many examples. It 
may be noted that though the Roman shaft at Wickham 
is straight-sided, there are other Roman examples of the 
kind with a distinct bellying. One of these is shown in 
Fig. 5, ante, p. 10, and another in the illustrations to 
Mr. Fox's paper on Uriconium, in the Archaeological Journal 
for 1897. In all the Monkwearmouth- Jarrow examples the 
mouldings are very delicately worked in the lathe, and give a 
high idea of the skill and industry of the craftsman of the time. 
As regards the use of them, it has been suggested that they 
formed part of the choir enclosures in the churches where they 
have been found ; but, on the other hand, we have found 
some still in their original position at Monkwearmouth, in- 
serted in the jambs of the window openings in the west wall 
of the nave. If this were the way they were used, the number 
found at the two churches would pretty well correspond to the 
number of the windows. Shafts of this particular kind seem 
not to have been found anywhere else in our own country, 
save at Hart in county Durham,^ nor have examples come to 
light on the Continent. The writer has searched in vain for 
their prototypes in Italy and Northern Gaul, the two sources 
from which the builders of these churches by Wear and Tyne 
are supposed to have drawn their inspiration. Anything less 

^ The Reliquary, January, 1894, p. 2. 



Fig. 1 1 1. — Fragments of baluster shafts 
in the museum at Dover. 

like them than the bold though clumsy carving of Gallo- 

Roman and Merovingian 
origin cannot well be imagined. 
Continental parallels to this 
earliest type of our Saxon 
shafts are apparently not to 
be found, though the later 
types of Saxon balusters to 
which we will now proceed 
can be more or less mated 

Passing away from this 
specially interesting northern 
group, we find a set wrought 
on a different design but with 
equal care, at the opposite 
extremity of the country in 
the museum at Dover, whither 

they have been transferred from the Saxon church of 

St. Mary-in-the-Castle (Fig. iii). They are in a very 

fragmentary condition, and some pieces 

have been worked into Early English 

mouldings. They have been turned in 

a lathe, and the forms are well empha- 
sized and the cutting sharp and clean. 

One example is remarkable because 

the shaft springs from a square plinth 

cut out of the same piece of stone 

with itself. These shafts have more 

of the normal baluster form, with a 

definite swelling in the middle part ; 

and this shape is still more distinctly Fig. 112.— Baluster shaft 

„ • ,1 1 r T> ^ from Barton-on-H umber. 

seen m the example from Barton-on- 

Humber (Fig. 112), where the profile is not unlike that of 

the little decorative baluster shafts in the north shown in 



Figs. 106, 108, 109. There is a well-known series of turned 
baluster shafts in the triforium openings of the transepts of 
St. Alban, figured in Buckler's work on the abbey, that are 
supposed to be part of the material collected for the rebuilding 
of the abbey church by abbot Ealdred at the end of the tenth 
century. This may be the truth, though it is worthy of note 
that they are of precisely the same stone, a fine grained oolite, 
as the Early Norman capitals and plinths and the plain round 



Fig. 1 1 3. — Shafts from the triforium of the Abbey Church of St. Alban. 

or octagonal shafts that occur in the same range of openings. 
The whole of the stone may, of course, have been taken from 
the ruins of Verulamium, but in this case one would expect to 
see some trace of a Roman tool on some of the stones, showing 
that they had been re-used. A careful examination of the 
shafts has revealed no indication of the kind. It may be noted 
here that the shafts, which are eight in number, are in no 
instance all in one piece, but are made up of short lengths 
averaging about 30 in., which are joined as shown in the 
examples chosen for illustration (Fig. 113). Some of the 



mouldings, especially at the bases, are made up with 
plaster, and the whole work presents a somewhat makeshift 
aspect, quite consistent with the theory that the pieces were 
survivals from an old store of building material. The caps on 
them are Norman, of a style like the one specimen shown. 

It is to be remarked that baluster shafts, sometimes banded, 
sometimes with swelling outlines, occur in early Romanesque 
work on the Continent, as well as occasionally in Norman 
buildings in England. The continental examples are to be 
found rather in the western than in the eastern of the two 
provinces already spoken of (ante, p. 48 f.). In Germany, 
partly owing to the influence of Italy, the shafts are as a 
rule of classical plainness and classical contour, but in the 

Loire district banded balusters of 
the tenth or eleventh century are 
to be found, as on the facade of 
the early church of St. Mexme 
at Chinon. The west front of 
Tewkesbury, the west end of 
Lindisfarne, and the tower of New- 
haven, Sussex, may be quoted as 
supplying Anglo-Norman parallels. 
Another class of shafts is repre- 
sented by the Northamptonshire 
examples illustrated in Figs. 114 
and 115. The first is one of the 
two shafts dividing the triple 
opening from the first story of 
the tower to the nave at the 
western end of Brixworth church, 
and the latter is one of many 
shafts, all of the same character, used in the tower of Earls 
Barton. These examples differ from those previously noticed 
in that they are not turned in the lathe, but are roughly hewn 
to shape by mallet and chisel. They may be regarded as 

Fig. 114. — Baluster shaft in 
opening in western wall 
of Brixworth Church. 



clumsy imitations of the turned balusters, the use of which 
begins, at any rate, at a much earlier period than that to 
which these Northamptonshire examples can be ascribed. The 
Brixworth shafts possess a cap and base of the normal cubical 
form, cut in the same piece with the shaft. 

Fig. 1 1 5. — Baluster from 
Earls Barton. 

Fig. 116. — Mid-wall baluster shaft from 
Wing, Bucks. 

With the above may be compared the similar but smaller 
mid-wall shaft (Fig. 116) that divides the opening high up 
in the east wall of the church of Wing, Bucks. This is a 
building which, on the evidence of its plan, has been attributed 
to the same age as the main structure at Brixworth. The 
opening at Brixworth containing the shafts (Fig. 114) is 
shown by distinct marks to be a later insertion in the wall 
where it is found. Here, at Wing, there are no such indica- 
tions, and if the window with its shaft be contemporary with 
the rest of the fabric it becomes a document of capital 
importance as bearing upon its date. In this connection it 
must be noticed that the imposts of superimposed slabs shown 
in Fig. 116 are of the same make as the imposts of the nave 
arcades in the church below. See postea, pp. 259, 268. 

The last examples selected are shafts of a simple bellying 
form without mouldings. Fig. 1 1 7 has a rude cap and base, 


and occurs in the very interesting tower of Bardsey church, near 
Leeds, noticed ante, p. 156. In this tall and slender tower (it 
only measures 12 ft. i in. in width on its western face but is 
nearly 50 ft. high) there are on the south side two stories of 
double openings with mid-wall shafts and through-stones, from 

one of which the illustration is taken. 
Finally, we may refer back to page 
92, and to Fig. 39, which is from 
the church of Worth, Sussex, where 
we find mid-wall shafts of rude work- 
manship dividing the double window 
openings on both sides of the nave. 
The church is of Late Saxon date, but 

the shape of the baluster, a plain 
Fig. 117. — Shaft from in- , r • 1 1 i- 

„ ' r J bellvmg shart without mouldings, is 

Bardsey near Leeds. . . . 

curiously like one particular form of 
Roman shaft referred to ante, p. 9, of which a few examples 
have been found. The Roman stones in question are un- 
moulded, and are not unlike the supports of floors in chambers 
with hypocaust arrangements. One at Chesters, on the North 
Tyne, measures 2 ft. 2 in. in height, and swells from a diameter 
at each end of about 5 in. to a middle thickness of nearly 
9 in. In the case of Worth there can be no question of a 
borrowing of Roman detail, as the church is in the middle 
of the Andredesweald and away from Roman sites. The 
resemblance is merely a coincidence, and does not help us in 
working out the problem of the relation to classical models 
of these interesting features of our pre-Conquest churches. 

Next to Earls Barton the most important in an historical 
sense of the pre-Conquest towers in England is that of Sompting 
in Sussex, for this has preserved its ancient finish shown in 
Fig. 90, ante, p. 163. The quoins which are partly in long- 
and-short work are cut back evenly along a vertical line and 
the plaster which covers the rubble work is brought up to this. 



A pilaster strip, half-round in section, runs up the middle of 
each face, and a little under the belfry openings a carved 
capital is introduced, though the strip does not end at it. 
The horizontal string course, a third of the way up, is cut 

Fig. 1 1 8. — Impost and cap of soffit shaft from southern jamb of tower arch, 

Sompting, Sussex. 

into a kind of billet moulding. At the summit of the tower 
each of the four walls ends in a gable, and the roof is formed 
by four planes lying on these and meeting in ridges above 
the apex of each gable. Those who are familiar with German 
churches will readily recognize the ' Helm ' form, which is 
common along the Rhine as well as in other parts. 



Internally a remarkable tower arch fixes the attention. It 
possesses the comparatively advanced feature of a soffit shaft 
and half-round roll moulding under the arch (Fig. ii8). The 
soffit shaft has a carved cap in the ornamentation of which the 
originally classical motive of leaves upright and turned over at 
the tips is easily to be discerned. The capital is really debased 
Corinthian and is the Anglo-Saxon counterpart of the corinthian 
caps already spoken of in connection with some early Roman- 
esque churches of Ger- 
many.i The impost, 
carved with bold vol- 
utes,' in the eye of 
which is a round mass 
of what looks like 
the seeds of a sun- 
flower, are returned 
along the face of the 
wall. Mounting to 
the belfry stage we 
find on each of the 
northern and southern 
faces of the tower two 
double openings with 
mid-wall shafts, and on each of the eastern and western 
two single openings with triangular heads. This duplication 
of the openings is rendered necessary by the presence of the 
upright pilaster strip, and this is motived by the form of the 
roof, so that the tower must have been designed as a whole 
as we have it. In these double belfry openings we find 
examples of the corbel caps above referred to. Three of 
them are of the form shown in Fig. 26, No. vi, ante p. 64, 
but that in the north-western double opening is of the simpler 
shape of B in Fig. 119, where it is placed in juxtaposition with 
the corbel cap A in the west front at Trier on the Moselle 

^ante, p. d"]. 






Fig. 119. — Corbel caps. 

Corbel cap at Trier. 

Base of shaft. 

Corbel cap at Sompting. 

Front view of top of cap. 

Grating that fills in the window opening. 



shown already in Figs. 25 and 26. The resemblance here is 
almost exact, though the Sompting cap is comparatively rude 
in execution, and it bears out the German character of the 
external termination of the tower. 

There are two fine Saxon towers that resemble Earls Barton 
in the possession of long-and-short quoins and a profusion of 
pilaster strips, and which introduce 
us to some new and interesting 
features. These towers are Barnack 
in Northamptonshire and Barton-on- 
Humber in the northernmost part of 
the county of .Lincoln. Barnack has 
unfortunately lost its belfry stage and 
is now crowned with a good Early 
English spire. It exhibits in the 
Saxon work that remains various 
openings which resemble more or 
less those at Earls Barton, while there 
are also some interesting pieces of 
decorative carving on the exterior, 
which make the tower of special im- 
portance. It is not proposed here Fig. 120.— Pierced mid-wall 
to enter upon the subject of this 
carved enrichment, for this cannot be 
treated properly except in connection 
with the stone crosses the consideration of which is at present 
reserved ; a word may however be said on the ornamental 
treatment of the mid-wall slabs which close the aperture of two 
of the smaller tower windows. These fillings have already 
come before us in the form of stone slabs with the aper- 
tures for light cut in them in a cruciform scheme (Earls 
Barton, East Lexham), or in keyhole shape (Langford in 
Oxfordshire). Here at Barnack the apertures are cut in a 
design of interlaced circles (Fig. 120), It is interesting to 
note that scanty remains of a wooden mid-wall slab with 

slab at Barnack, North- 
ants. From a drawing 
by J. T. Irvine. 


the apertures similarly fashioned are to be seen in a Saxon 
window discovered in 1869 in the north wall of the chancel 
of Birstall church near Leicester. Other wooden mid-wall 
slabs with apertures not forming an ornamental pattern are 
to be seen in situ at Barton-on-Humber ; Houghton-on-the- 
Hill, Norfolk ; South Lopham, Suffolk ; Stevington, Bedford- 

The slabs in these examples have been found embedded in 
the centre of the thickness of the walls and are evidently 
contemporary specimens of the woodwork used by Saxon 

The ground story at Barnack is well worthy of attention, 
but as it leads us on necessarily to speak of some new types 
of Saxon structures, it will be well to reserve it for consideration 
in another chapter. 



Vlll. The Tower forming the Body of a Church. 

The ground story of the tower at Barnack provides a fitting 
introduction to an interesting type of Saxon church in which 
such a ground story forms the chief space of the interior. At 
Barnack the ground story was 
evidently used for some purpose 
of dignity, though not necessarily 
an ecclesiastical one, and it opens 
towards the church by a tower 
arch of exceptional width, so that it 
commands the whole interior east- 
wards. Ingress from the exterior 
is gained through a lateral door- 
way, a rare feature as we have 
seen in Saxon towers. The plan, 
Fig. 121, and view into the interior 
of the tower from the east through the tower arch. Fig. 122, 
will give an idea of what the ground story of Barnack has to 

The span of the tower arch is close upon 13 ft. and the 
jambs are 1 5 ft. high to the top of the imposts. These are 
very curious in their form. \Yhei"eas at Wittering (Fig. 59, 
ante, p. 108) and Market Overton (Fig. 46, ante, p. 99) they 

Fig. 121. — Barnack Tower. 


are huge and simple trapezoidal blocks, we find them here, 
though in stone, carved into what appears the similitude of 


FiG, 12 2. — Tower arch, Barnack. 

superimposed Roman tiles. Imposts actually formed in the 
latter fashion occur in many Saxon buildings, such as St. 
Pancras, Canterbury ; Brixworth ; and Wing, Bucks, and the 
Barnack carver seems to have had such models in his mind 


when he set himself to hew his fine local material into shapes 
in which there is so little stone character. 

A feature of unique interest is to be seen in the western 
wall of the tower. This is a recess apparently intended for a 
seat, with a triangular head starting from square imposts that 
surmount the jambs. It is about 3 ft. 6 in. in width, by 
a depth of i ft. 4 in. The seat is i ft. 8 in. from the 
present floor. When in 1854-5 the tower was cleared out 
and restored under the care of the late Canon Argles, some 
indications were found of a range of seats topped with wood 
that ran round the tower walls on each side of this central 
throne. In the north and south walls near to the tower arch 
and 4 ft. from the ground, there are small niches, resem- 
bling aumbries, like those in the upper chamber of the tower at 
Deerhurst, mentioned ante, p. 168. 

It will be remembered that the Saxon church or part of it 
was a recognized place for legal transactions (see vol. r, p. 
371) and we have the definite statement that the south porch 
of the cathedral at Canterbury was used as a court of justice. 
It is quite in accordance with likelihood to imagine the spacious 
interior of the tower at Barnack employed for a similar pur- 
pose. The presiding official has his place marked out for him 
in the western niche while other persons charged with the 
conduct of business would be accommodated on the lateral 
benches. The wide tower arch would enable the people gener- 
ally to attend the proceedings without crowding into the space 
reserved for the officials.^ 

^ It has been argued that these aumbry-like niches at Barnack have an 
ecclesiastical significance, and that their presence is inconsistent with the 
theory of the use of the ground story of the tower for legal business. But 
the recesses in Deerhurst tower and in that of Skipwith, Fig. 93, ante, 
p. 168, were not necessarily connected with altars. Moreover even if the 
recesses at Barnack were altar-aumbries this would not preclude the use of 
the tower for placita. There was an altar to St. Gregory in the porch-tower 
at Canterbury that was actually so used. See Willis, Canterbury Cathedral^ 
p. I I. F"or Barnack see Journal of Arch. Ass., 1899. 


The upper stages of the tower present the usual appearance 
of former habitation, and Mr. Irvine, who devoted much 
attention to Barnack, worked out a scheme of arrangements 
by which the tower could have been made fit for residential 
purposes. From the position of the windows he argued that 
the principal apartment, now the ringing chamber, was divided 
into two with an intermediate passage, while there were other 
chambers above. The doorway, which here as elsewhere opens 
in the east wall of the tower, is at the level of nearly 35 ft. above 
the ground, and gave access in his view to a space between the 
flat roof of the nave and the external gable, the mark of which, 
about 12 ft. higher up, is to be seen on the east face of the 
tower. In connection with this question of residence it must 
be noted that the level of the supposed principal apartments 
was more than 30 ft. above the floor of the tower, and even 
allowing, as Mr. Irvine assumed, an intermediate landing, the 
access by ladders must have been very troublesome, and the 
whole residence an extremely inconvenient one for persons 
of any distinction. 

The tower of Barton-on-Humber as shown in Fig. 123 
exhibits long-and-short quoins and pilaster strips which are 
joined by round and triangular arches in a fashion rather more 
logical than that illustrated at Earls Barton. On the ground 
story of the tower there are doorways on the south and north, 
a tower arch opening into the nave of the church, and another 
archway giving access on the west to an adjunct of Saxon date 
and similar workmanship, but without the pilaster strips, which 
was probably contemporaneous with the tower. In this 
western adjunct there are the marks of a wide western door- 
way now blocked, but this looks comparatively modern, and 
it is uncertain therefore whether or not a narrower Saxon door 
once existed in the same position. 

The windows in the adjunct consist in a round-headed 
double-splayed light on each lateral face and two circular 
double-splayed openings one above the other in the western 




wal]. In both of these last there still remain portions of the 
original wooden mid-wall slab, pierced with a number of holes 
three-quarters of an inch in diameter for the transmission of 

The tower shows on its first story on each lateral face a 
double window, round-headed, with mid-wall shaft that is here 
of a distinct baluster form as 
shown in Fig. 112, ante, p. 196. 
Higher up on each of the four 
faces is another siich window 
with a triangular head. The 
uppermost story contains 
double belfry openings divided 
by mid-wall shafts that have 
no baluster-like form, with 
curious capitals of which No. x i 
in Fig. 99, ante, p. 180, gives 
a specimen, and through-stones 
exactly of the type we have 
become familiar with in the 
towers of the ' Lincolnshire ' 
group. This stage is later in 
date than the rest of the tower, 
but like the other parts just 
described is of Saxon work- 
manship. The handsome and 

very spacious aisled nave and chancel are of the later mediaeval 

The point of chief interest for our purpose is the evidence 
that the building affords of a type of early church of a some- 
what singular kind, in which the ground story of a square 
tower forms the nave or body of the oratory, a small chancel 
being built on to the east of it. In the year 1898 some 
investigations kindly made by Mr. Hodgson Fowler, in the 
course of extensive works he was in charge of at Barton-on- 

FiG. 124. — Western wall of nave, 
Barton-on-Humber Church, show- 
ing marks of original chancel walls 
of Saxon date. 




Humber church, brought to light direct evidence that in its 
earliest condition, or at any rate in Saxon days, the tower 
with its western adjunct was the church, and nothing appeared 
east of this save a small square-ended presbytery some fifteen 
feet in internal length. The proof of this is worth giving. 
The eastern wall of the tower, forming the western wall of the 
existing nave, was stripped of its plastering in 1897 and 
disclosed the distinct marks of side-walls projecting from it 
to the east (Fig. 124), That these were the walls of the 
original chancel was proved by excavation which laid bare 

^^p **T=- 

Fig. 125. — Plan of original Saxon church at Barton-on-Humber. 

the south-east quoin and enough of the foundations to show 
its size and shape.^ The result is the ground-plan given in 
Fig. 125, while Fig. 126 when compared with Fig. 123 
exhibits the contrast between the aspect of the original Saxon 
church of which the ground story of the tower was the nave, 
and the later mediaeval structure in which the original small 
chancel has developed into a relatively enormous edifice. The 
uppermost story of the Saxon tower has been replaced in 
Fig. 126 by a saddle-back roof, for which there seems some 
evidence at Corbridge, and the gables are finished with stone 

^ The success of these investigations was largely due to the interest shown in 
the archaeological questions involved by Mr. John Briggs of Barton-on-Humber, 
who was carrying out the work at the church, and was at the pains to make 
the necessary excavations. 
















crosses after the fashion of one preserved in Corbridge 
church, and shown in Fig. 127. 

This is a striking object lesson in the growth of the 
mediaeval church. Such growth in nearly every case resulted 
in a great proportionate extension of the eastern part. It is 
rare to find a Saxon chancel complete, that portion of the 
church having in most cases been amplified in later times. 
At Bosham, Sussex, there are distinct marks of two exten- 
sions, in the Early Norman and in the 
Early English styles. This is only 
in accordance with the tendency which 
in more advanced mediaeval days so 
modified the interiors of some of our 
greater churches that the eastern part 
became a complete church of cruciform 
plan, the nave passing out of impor- 
tance. This is what took place at 
Canterbury, Lincoln, and elsewhere. 
Here at Barton-on-H umber it was not 
that the chancel was enlarged, but a 
whole new church, nave and aisles 

and chancel together, was substituted for it, the tower and 
western adjunct being finally relegated to the condition of 
lumber sheds. 

Not far away, at Broughton near Brigg in Lincolnshire, 
another example of this same type of church can be detected. 
The church of this village, noteworthy as one of the only two 
which even touch the straight Roman road from Lincoln to 


the Humber in its course of thirty miles, has a square western 
tower with a later semi-circular stair turret on its western face. 
The present tower arch is ornate on the side towards the 
tower but very plain towards the church, and is clearly the 
original chancel arch of a church of the form now fixed at 
the neighbouring Barton-on-Humber. Some indications of 
a small chancel were found here several years ago in 

Fig. 127. — Saxon gable 
cross from Corbridge 


connection with works for the heating apparatus.^ It needs 
however only a glance at the drawings of the two faces of the 
tower arch (Figs. 128 and 129) to see that the western face 
was the one intended to be seen. The same is the case with 
the tower arch at Barton-on-Humber, which though not as 
fine as the arch at Broughton has some effort after enrichment 
on the western side, the eastern face being quite plain, see 
Figs. 124 and 130. 

The Broughton arch has the curious peculiarity that the 
angle shaft on the north side in its lower part has been almost 
worn away, apparently by the process of using it as a whetstone 
for the sharpening of tools or implements. Such marks, 
generally in the form of narrow cuts or grooves, are often 
enough seen on the quoin stones or on the porches of our 
ancient churches. There are many on the porch at Good- 
manham, Yorkshire. They have sometimes been explained 
as the marks made by sharpening arrow-heads in the good old 
days of the long-bow, when the butts were set up near the 
church, and the graveyard was used as a place of muster if not 
of actual exercise (see vol. i, p. 364). In most cases, how- 
ever, as here at Broughton, the author of the mysterious 
indications is more likely to have been some gravedigger or 
gardener of the seventeenth or eighteenth centuries. 

We have thus at Barton and Broughton two certain instances 
of a type of church in which the tower formed what we should 
call the nave or main division of the interior. It is a type 
for which a parallel may be found in Germany at Werden 
a. d. Ruhr, where, according to Effmann's demonstration, a 
structure of the kind was reared about the year 900.^ It 
would be interesting to know whether among our Anglo- 
Saxon buildings there are other examples that can be included 
in the same category. The difference between a tower thus 
treated and an ordinary western tower will be seen in the 

^ See Archaeolo^cal Journal^ mi, 335. 

'' Die Karol. Otton. Bauien zu Werden, p. 168. 













eastern quoins. When the tower forms the body of the 
church the chancel which is joined to it will be narrower than 
the tower, and this will accordingly have its eastern quoins 
complete to the ground as they show equally with those at 
the west. The normal western tower, as in the case of all 
those of the ' Lincolnshire ' type, is narrower than the nave to 
which it is joined. The eastern wall of the tower is part of 
the western wall of the church and this projects for one or 
two feet on each side showing its own quoins. Hence the 
appearance of nave quoins on each side of a tower removes it 
out of the category now under discussion, while on the other 
hand when the eastern tower quoins come down to the ground 
there is a possibility that it once had a small chancel to the 
east of it. Earls Barton answers to this description, whereas 
Barnacle, the aspect of which would suggest a tower church, 
exhibits traces of the quoins of an old nave to north and 
south, and must have been all along a western tower. 

The type of church which has just been discussed presents 
some difficulties in the matter of lighting and arrangement. 
Both Barton and Broughton show those doorways in the eastern 
face of the tower above the tower arch which have already 
formed the subject of discussion (ante, p. 169 f.). At Barton 
the north and south round-headed double windows with the 
mid-wall balusters are on about the same level as this doorway, 
and might conveniently have lighted a chamber on the first 
floor of the tower corresponding therewith, but in this 
case there would have been no opening for light in the 
ground story of the tower below the floor of this chamber, and 
the whole interior would have been dimly illumined through 
the lateral double-splayed lights of the western adjunct and the 
small chancel. The circular opening in the lower part of the 
west wall, with its mid-wall slab of wood pierced with holes, 
would have helped but little. On the other hand if we assume 
that there was no such floor in the tower, so that the interior 
would receive light through the baluster windows, what could 



be the use of the doorway above the tower arch in the east 
wall ? It is conceivable that it may have given access by 
means of a ladder from the ground story of the tower to a 
chamber contrived in the roof over the chancel, where the 
sacrist may have had his abode. The existence of such 
chambers over chancels has been already established and 
references have been given to parallels in Ireland and in 
Norman work in our own country, see ante, p. 171. The 

■^ ?.^J.?.T..'.H?.....^.!-.P.o?.. 

b f" TOW E R 


:30. — Section of tower and western adjunct, Barton-on-Humber. 

western adjunct at Barton may have been similarly treated, 
for there is an upper opening into it from the tower, opposite 
to that from the tower to the chancel, and the duplication of 
the circular lights in the west wall would suit such an 

In Fig. 130 is given a section of the tower and western 
adjunct with indication of the relative levels which will illustrate 
the foregoing. It will be seen that the western face of the 
tower (formerly chancel) arch is decoratively treated with 


projecting imposts, pilaster strips, and a corresponding hood 
mould, while a slab above this last, shown in section, is 
carved at the top with a human head in relief and was probably 
designed to carry a representation of the Crucified. The 
eastern face of the same arch is quite plain, see Fig. 124, 
ante, p. 209. It may be noticed in passing that the enrich- 
ment of the eastern face of the archway between the tower 
and the western adjunct, while the western face is plain, is a 
proof that the western adjunct is not later than the tower 
and is probably part of the original scheme. 

At Broughton there is a small window only 12 in. wide in 
each lateral face of the tower about 1 2 ft. 6 in. from its floor, 
which may have lighted, though somewhat dimly, the interior. 
Above these there may have come a floor, for the upper door- 
way in the east wall towards the present church is about 17 ft. 
above the ground level of the tower, and a floor corresponding 
with it would come above the side windows. There is no 
indication here of a western adjunct, but the three-quarter 
round turret on the western face is a later addition that may 
not have been there when the tower formed the primitive 
church. Broughton tower has a doorway in the southern wall 
quite at its western end, but no doorway on the north. The 
two lateral doorways at Barton-on-Humber are also towards 
the western limit of their walls, and appear like the single 
one at Broughton to be placed with a view to use by a 

IX. Axial Towers. 

Barton-on-Humber as we see it now is a large mediaeval 
church with a western tower that forms no part of the interior, 
and the discovery that this now abandoned space was in 
reality the original church is so surprising, that the arrange- 
ment seems at first sight more peculiar than it really is. It is 
at any rate an easy transition from towers used in old time as 


those of Barton and Broughton were used to the kind of tower 
here called axial. The axial tower is one in which the ground 
story forms one division of the church as it is traversed from 
west to east, and may be (i) situated at the western end, (2) 
between nave and chancel, or (3) at the eastern end where its 
ground story forms the chancel itself. It differs from the 
ordinary western tower, and from the tower that forms the 
body of the church, in that it is practically the same width as 
the body of the edifice, whereas the others are either narrower 
or wider, and it must be distinguished also from the central 
tower of a cruciform plan where the tower is flanked by 
transepts or transeptal chapels. 

(i) The fine church of Hooton Pagnell, near Doncaster, 
possesses now a western tower, a nave with north aisle, and an 
extensive chancel. The western third of the last, the nave, and 
the ground story of the tower, are built in rubble work that 
might well be of Saxon date, and this work extends on the 
tower to a height a little greater than that of the nave walls. 
The recessed tower arch and the chancel arch are original, and 
the latter has early detail. The construction of the arches 
however hardly looks to be Saxon and an Early Norman date 
for the church seems most probable. The point of interest for 
the present purpose is that the tower is externally the same 
width as the nave but the wall of it is thicker, so that the 
internal measurement of the tower is less than that of the body 
of the church. This thickening of the walls, with the existing 
tower arch and the evidence of the exterior masonry, make it 
probable that a tower was contemplated in the original plan of 
the church, and this leads to the inquiry whether other early 
churches, which have a western division with thickened walls, 
were intended originally to be finished with a western axial 
tower like that at Hooton Pagnell.^ 

'The Hampshire example of Boarhunt, Fig. 55, ante, p. 105, exhibits 
the marks of a western division without any thickening of the wall at 
compared with that of the nave. 



Diddlebury, Shropshire, may be adduced as actually ex- 
hibiting a tower of the kind, but Diddlebury, an interesting 
monument, needs some analysis. The earliest part is the 
north wall of the nave, with a small portion of the return wall 
forming the western limit of the nave and dividing it from the 
western tower, which is the same width as the nave. That this 
earliest part of the fabric is Saxon is proved by the very 

characteristic north doorway and 

double-splayed window shown in 

Figs. 44 and 40, ante, pp. 95, 93. 

The wall in which these features 

appear is treated quite exceptionally. 

On the exterior it exhibits squared 

ashlar work of a quality almost 

equal to that at Bradford-on-Avon ; 

the courses being of irregular heights 

Fig. 131. — Herring-bone facing but carried in each case consistently 

from the interior of Diddle- ^^^^g ^.^g wall, and showing careful 

bury Church, Shropshire. ^^^j^^ ^^ ^j^^ ^^^^^^^ ^^j^ -^ 

original work, for the jambs of the double-splayed window are 
included in it. In the interior on the other hand, there is a 
facing of herring-bone work in which the stones are carefully 
cut to shape and fitted, with mortar joints ^ to ^ in. in thick- 
ness. See Fig. 131. 

Along the foot of the north wall on the exterior runs a plinth 
of three square orders, and this continues without a break from 
the original eastern quoin of the Saxon nave to the present 
north-western quoin of the western tower. If the tower wall 
above this Saxon plinth were also Saxon, we should have an 
undoubted Saxon western axial tower, but unfortunately the 
tower wall in its oldest part, the eastern half of the northern 
face, is not sufficiently like the north wall of the nave for us 
to accept it as pre-Conquest. The original thickness of the 
tower wall either on the east or on the north cannot now be 
ascertained. Most of the tower is of Norman and later dates, 


and it is encumbered with unwieldy later buttresses, one of 
which hides the junction of nave and tower. 

With Diddlebury may be mentioned Daglingworth, Glou- 
cestershire, where the Saxon nave had a square western 
division of which the wall was thicker by more than a foot 
than that of the nave. The wall between this and the nave 
has been removed, but it is known that it opened towards 
the nave by a wide arch. 

In connection with these western divisions, this will be a 
fitting place to introduce a notice of two buildings of remark- 
able interest in East Anglia, one of which has been known 
and discussed for some time past, while the other has had 
its ecclesiastical character and early date fixed by investigations 
recently carried out.^ 

Each of the structures in question is located at a place called 
Elmham. There is a North and a South Elmham in East 
Anglia, the former in the heart of Norfolk a few miles north 
of East Dereham, the latter in the north eastern corner of 
Suffolk not far from Bungay. Now when Theodore of 
Canterbury, about the year 673, had divided the East Anglian 
diocese the original seat of which was Dunwich on the Suffolk 
coast, we find the second bishop located at ' Elmham.' This 
was obviously the Norfolk Elmham, for South Elmham in 
Suffolk is only a dozen miles from Dunwich, and it is simply 
not arguable that the second seat of the divided diocese can 
have been located so close to the parent see. The succession 
of the East Anglian bishops was interrupted for a hundred 
years in the Danish period, but in 956 the see was re-established 
at * Elmham ' as the one East Anglian bishopric- At that 
epoch a bishop's church must have been built or restored, and 
this church may have been renewed at any time before the 
East Anglian see was removed in 1075 ^^ Thetford, to be 
finally located in 1094 at Norwich. 

^ See The Builder^ vol. lxxxiv, 1903. 

^ See Stubbs, Registrum Sacrum Anglicanum, Oxford, 1897, p. 230 f. 


Which of the two Elmhams was the seat of the restored 
bishopric of 956 is another question. North Elmham was 
after the Conquest in the possession of the see of Norfolk 
and the bishop had a residence there, but South Elmham and 
its neighbourhood were also associated with the see, and we 
have the following curious notice in Domesday about Hoxne 
in Suffolk, some nine miles from South Elmham, ' in h man. e 
^cclia sedes episcopatus de Sudfolc. T.R.E.' ^ At both the 
Elmhams there exist ruined buildings which on their face- 
values would be ascribed to the eleventh or twelfth century, 
but which may be in fact earlier than their general aspect seems 
to suggest. The building at South Elmham in Suffolk is 
called the ' Old Minster ' and there is a description of it by 
Mr. C. E. Peers in the Archaeological Journal^ vol. lviii, 
p. 423 f. Fig. 132 gives the plan. 

The structure which is of flint rubble exhibits little, save 
its size and some peculiarities of plan, to mark it off from other 
Norman apsidal chapels in the same region, such as that of 
Mells near Blythborough, or the chapel outside the keep at 
Castle Rising. In material, technique, and form of window 
openings it accords with a Norman origin. The chief 
peculiarity, over and above an abnormally wide chancel arch, is 
the existence of a division in the nave cutting off a western 
portion the walls of which are thicker by about 8 in. than 
the walls further east, though flush with these on the exterior. 
The difference is rather more than that observed at Hooton 
Pagnell, where the western division as we have seen now 
carries a tower. Whether a tower existed at South Elmham 
cannot be absolutely determined. Apart from the thickening 
of the wall there are no appearances of this, but it is a 
possibility of which account should be taken. The fact that 
the partition wall across the nave is no mere screen but a 
wall as thick as the external ones to the west of it, is in 
favour of a tower. It is against it on the other hand that 

^11. 379- 


there is no single tower arch but two openings into the nave ; 
while the similarity of the fenestration in the western division 
and in the nave also militates against the supposition of a 
tower. What other reason however can there be for the 
thickening of the wall ? 

At North Elmham there exists also a ruined church, of 
which Fig. 133 gives the plan. This it will be seen at once pre- 
sents features of novelty and interest. There is a long narrow 
aisleless nave ending in a transept, in the east wall of which is 
the arch of the presbytery opening into an apse. The plan is 
thus an example of the ' crux commissa ' about which a word 
will be said on a coming page. West of the transept, and 
opening into the nave through doorways, are two chapels.^ A 
curious technical peculiarity is to be observed in the filling of 
the re-entrant angles on the plan with buttress-like pieces of 
the shape of quarter cylinders. These are to all appearance 
of contemporary date with the walling. For the purpose on 
hand we are chiefly concerned with the western end. Here is 
a part divided off from the rest of the nave, with which it 
communicates through one opening 10 ft. wide instead of by 
two openings as at South Elmham. The walls of this 
sundered portion are about 5 ft. thick, whereas the walls 
generally measure 3 ft. 3 in. This fact in itself suggests a 
tower, and the presumption is raised almost to certainty by the 
existence at its south-eastern corner of a projecting turret half- 
round in plan containing the remains of a spiral stair. Such a 
stair turret on the western face of a tower occurs, we have 
seen, in four examples of Saxon date (ante, p. 175). Here 
at North Elmham the position suggests that of the flanking 
stair turrets so common in German work, but the feature is 
single and not double. 

All that has been attempted here is to present these two 

* The northern chapel has been much pulled about, and in its external 
outline it may originally have corresponded with that upon the south. The 
whole building is encumbered by additions and alterations of later date. 














Elmham ruins as examples of the particular type of plan here 
under notice. No discussion of their date can be satisfactory 
without a proper historical treatment of the whole subject of 
the two Elmhams and their relations to the East Anglian sees, 
and this it is hoped may before long be accomplished by a 
competent authority on East Anglian lore. If a pre-Norman 
date be indicated for the N. Elmham ruin, it is not because 
any specifically Saxon characteristics are to be discerned in it, 
but because the earthworks of an Early Norman burh, or 
moated mound, almost overlie its western end and in all likeli- 
hood are posterior to it in date. 

There is accordingly some reason to regard the axial western 
tower as a Saxon form perpetuated in England in Norman 
times. The western adjuncts (Barton-on-Humber) and the 
western divisions in naves (Boarhunt, Daglingworth, Diddle- 
bury) must be taken into account as indicating stages in the 
evolution of the scheme. It is worth notice that in German 
Romanesque there are some examples of the scheme, and 
one or two are indicated below.^ 

(2) The Axial tower between nave and chancel we find at 
Dunham Magna in Norfolk, a Late Saxon building of much 
interest. The triangular-headed western door, now blocked, is 
seen in the view Fig. 134, and the shallow arcading which runs 
round the wall of the nave in the interior has been noticed in 
connection with the similar feature on the exterior of Bradford- 
on-Avon (Fig. 77, ante, p. 137). The tower with its belfry- 
openings is of the plain 'Lincolnshire' type but possesses 
long-and-short quoins. The plan of the church is given in 
Fig. 135, but the chancel, which is said to have ended apsidally,^ 
is of later work. It is to be noticed that the north and south 
walls of the tower are on the exterior flush with the walls 
of the nave but they are thicker, so that the interior space 

^ Clemen, Die Kunstdenkm'dler der Rkeinprovinz, i, 4, Bedburg by Cleve ; 
III, 2, Gruiten by Elberfeld ; iv, 4, Diirscheven by Euskirchen. 
2 Archaeological Institute, 'Norwich' volume, 185 1, p. 216, 

cr -c! 









under the tower is less in width than the nave by about two 
feet. The eastern quoins of the tower come clear to the 
ground but the western ones stop when they reach the summit 
of the walls of the nave. 

Besides Dunham Magna there is no other church with axial 
tower between nave and chancel that is so completely Saxon, 
but hard by at Newton, close to Castle Acre, there is another 
axial tower of Saxon date that stands between a later nave and 
chancel, and at Langford in Oxfordshire a very fine church of 
post-Conquest date enshrines a Saxon axial tower that has features 

Fig. 135. — Plan of Dunham Magna, Norfolk. 

worthy of remark. Its double-splayed windows, in which the 
aperture is cut in a keyhole shape in a mid-wall slab, have been 
already mentioned (ante, p. 166). It has external pilaster 
strips that start and end with a step-like finish that reminds us 
of the similar features at Bradford-on-Avon, and are of the 
abnormal width for Saxon work of a little over a foot. Waith 
near Grimsby, Lincolnshire, and Northleigh, Oxfordshire, have 
Saxon axial towers in later setting, and the same is probably 
true of Guildford, Surrey. Here a massive tower with double- 
splayed lights and pilaster strips is embosomed in later struc- 
tures, but as the strips come down to the ground on the north 
and south sides where they are visible now inside the church, 
it is clear that these sides of the tower were once external. 
(3) The ground story of a tower used as a chancel, a rarer 

plan thaii the one just noted, occurs in Saxon work at 
II p 


Weybourn in the north of Norfolk, and it is said that the tower 
of Godalming church, Surrey, before the modern alterations, 
bore clear evidence that it was once the chancel of the earlier 

The Saxon tower of St. Peter, Bedford, occupies a rather 
curious position. It is now axial, but the chancel which it 
precedes is of greater width by about a foot than the tower. 
The chancel may conceivably have been the original nave and 
the tower a western one, while there are signs at the western 
face of the tower that there was once some building joined to 
it in this part. The eastern face of the tower possessed a 
triangular-headed opening on the first story corresponding to 
the openings already studied in the eastern walls of western 
towers above tower arches, and this fact lends force to the 
suggestion just made. The opening is visible now in the 
interior of the tower. 

X. Central Towers, 
Transepts, and the Cruciform Plan. 

We possess one complete Saxon cruciform church with central 
tower, two others of which the plan is clear though parts of 
the church are no longer Saxon, and one or two more that are 
not completely developed or about the form of which there is 
some obscurity. St. Mary, Dover Castle, is Saxon throughout 
(see Fig. 171, postea, p. 307) ; Stow, Lincolnshire, has Saxon 
transepts, Norman nave and chancel, and later central tower 
replacing a Saxon one ; while Norton, Durham, gives us a 
Saxon tower and north transept with part of a south transept, 
and a nave and chancel that are mainly of Transitional and 
Early English date. The other buildings are Breamore, 
Hants ; Deerhurst, Gloucestershire ; Wootton Wawen, War- 
wickshire ; Stanton Lacy, Shropshire ; Repton, Derbyshire ; 
Britford, Wilts ; Worth, Sussex, and the Saxon abbey church 


at Peterborough, the foundations of which exist under the 
present structure. Some of these, such as Worth, are com- 
pletely cruciform but have transepts lower than the nave and no 
central tower. Others like Breamore have the tower but only 
undeveloped transepts, and there is about all the others some 
uncertainty as to whether their original scheme included a 
central tower, or whether they were ever completely cruciform. 

We approach here a subject that involves no little com- 
plication and difficulty, and will treat (i) the growth of the 
side chapel or porch into the transept, (2) the central tower 
in its relation to the cruciform scheme, and (3) the so-called 
' crux commissa ' or T shaped plan. 

(i) The side chapel not a porch is found in its simplest form 
on the south side of the supposed original fabric of St. Martin 
by Canterbury and is indicated in Fig. 70, ante, p. 1 20. It also 
occurs at St. Pancras, Canterbury. The side porch that is at the 
same time a side chapel was a feature of the Saxon cathedral at 
Canterbury (postea, p. 260 f.) and we may probably take it to 
exist at Bradford-on-Avon and at Bishopstone, ante, p. 131 f. 

In the last mentioned case the projecting features occur at 
the western end of the nave and at Bradford they are west of 
the centre, while in the Canterbury examples they are almost 
exactly in the middle of the length of the churches. It remains 
to find such projecting features at the eastern ends. 

Now at Britford near Salisbury, there exists a substantially 
Saxon nave to which a later central tower, transepts, and chancel 
have been added. This nave, Fig. 136, measures 44 ft. 4 in. 
in length by a width of 20 ft. 2 in., and at the extreme eastern 
end of it there were found some years ago two very remarkable 
arched openings in the north and south walls. They had been 
built up and plastered over, and being now carefully cleared and 
protected by small outbuildings on the exterior, they appear in 
very good preservation. The northern archway is 5 ft. 9 in. 
wide by a height of 7 ft. 10 in., that on the south 5 ft. 7 in. 
wide and 7 ft. 8^ in. high. The present south doorway into 


the nave further west than these openings is in a third Saxon 
archway 8 ft. 9 in. high by 5 ft. 9 in. wide, but it is probable 
that this third opening has no special connection with the two 
others. These last correspond pretty closely in position and in 
size, but are curiously different in technical treatment. The arch 
of the southern opening (Fig. 137) is turned in large Roman 
bricks evidently re-used. Some of them are voussoir shape, 

Fig. 136. — Plan of Saxon portion of Britford Church by Salisbury. 
(The parts to the east of the openings are later.) 

about 13 in. long by a thickness of 3 in. at one end tapering 
to 2 in. They are not however all set voussoir fashion so 
that this wedge shape shall fit the form of the arch, but as 
often as not they are reversed so that the thin edge instead 
of the thick is on the extrados of the arch. The necessary 
wedse-like forms without which the arch could not be con- 
structed are given by the mortar joints, which are thicker 
on the extrados than below. 

The jambs are lined by tall and narrow upright stones, about 
4 ft. 6 in. high by 9 in. wide, standing on plinths and set at 
the outer thirds of the jamb with a recess in the interval 
between them, the whole thickness of the wall being 2 ft. 5 in. 



They are crowned by imposts which show the remarkable 
peculiarity already observed in Roman work and at Escomb 
(see ante, pp. 5, 114) that they are cut away to receive the head 
of the jamb stones which are mortised into them (Fig. 137). 
On the exterior face of the wall, now made conveniently 
accessible from the inside, there was a square-sectioned strip 
of stone 2| in. in face by a projection of i| in. that ascended 
the jamb and then followed the curve of the arch after the 

Fig. 137. — Face of jamb with springing 
of southern arch, Britford, Wilts. 

Fig. 138. — Soffit of northern arch, 
Britford, Wilts. 

manner of a hood mould. The imposts were probably returned 
along the outer face of the wall to meet this strip. The same 
feature occurs on the exterior face of the northern opening, 
and there are pretty clear indications on the inner side of the 
south opening that a similar strip had appeared on this face 
also. At the right-hand of the drawing in Fig. 137 is shown 
the vertical pilaster strip that, like the impost, has been hacked 
away flush with the wall and then covered with plaster, from 
which it has now been freed. The traces of this strip-work 
are of great chronological significance. 


The soffit of the northern opening (Fig. 138) is treated 
quite differently. It is panelled, so to say, with flat square 
slabs that are cut on their faces to the curve of the arch, 
and that leave between them recesses, like cassettes. The 
work is very careful, for the curved soffit slabs are framed 
as it were by bricks set edgeways, and bricks form the floor 
of the recesses or cassettes. 

The jambs have the plinth, imposts, and upright stones 
like the other archway, but these are not let into the imposts. 
In the space between the uprights there are square slabs 
with recesses above and below them. The most remarkable 
feature of the whole work is the ornamentation on the up- 
right jamb stones and intermediate squares on the eastern 
jamb of this northern opening. This carving is in its char- 
acter unique in Saxon architectural work, though it may be 
paralleled on the sculptured stones, in connection with which 
the writer hopes to deal with it on a future occasion. 

There are other details which might be noticed and which 
constitute differences between the two openings, but enough 
has now been said about their technical treatment. The 
purpose for which they were intended is a matter for 

There are practically three alternatives. They may have 
been (i) doorways to the exterior, (2) arcade openings, the 
survivors of a series giving access to side aisles, (3) archways 
admitting to side chapels, (i) is excluded, not because there 
is no rebate for doors, for Saxon doorways in most cases (ante, 
p. 103) appear not to have had rebates, but because the orna- 
mentation on the jambs is quite out of character with mere 
doorways. (2) The southern opening would work into the 
scheme of an arcade with the more westerly opening on the 
same side where is now the doorway of entrance, but the 
piers between the openings of such an arcade would have to 
be about 6 ft. wide. The arches however are too small in 
scale, especially too low, in proportion to the width of the 




o .S 


a. CO 

c ^ 

^ o 

(U ^ 

t^ '" 


O Oh 
P-1 rt 

1 -^ 

O u 



>-, _2 
3 Oh 

U c 

C X 
O rt 
4-. CO 

° -^ 

4-1 U 







nave for us to suppose them arcade openings. (3) There 
remains the supposition that they gave access at one time 
to side chapels, in which connection their ornate appearance 
would be quite in character, and their dimensions would be 
proportioned rather to the presumably small size of the chapels 
than to that of the nave out of which they led. Assuming 
this to have been their destination the eastward position of the 
side chapels is significant. 

The same arrangement meets us at Repton, Derbyshire, where 
is a Late Saxon chancel and beneath it a crypt that will occupy 
our attention later on. Besides the crypt and chancel the 
church has preserved relics of the eastern part of the Saxon 
nave with side chapels that measured internally about 15 ft. 
west to east by about 8 ft. north and south. There is a 
doubt as to the character of the access to them from the nave, 
but two Late Saxon columns are now preserved in the southern 
porch of the church which seem the relics of a set that were 
used as soffit shafts to archways that opened from the nave 
to these side spaces. One of these shafts is represented in Fig. 
156, postea, p. 259. At Deerhurst, Gloucestershire, at the 
eastern end of the nave, there were side chapels or as they 
have been termed choir aisles, about 12 ft. square internally, to 
which access was gained from the nave, on the ground floor 
at any rate, only through doorways. (See Figs. 139, 140.) 

The growth of such side chapels into transepts may be 
illustrated by the example at Breamore, Hants. This spacious 
and handsome church, Saxon throughout save for the south 
porch which is of later date, has only been recognized as such 
within the last few years. The plan is given in Fig. 141 and 
it must be noted that in the north wall there is the mark of 
the gable of a transept or side chapel that corresponded to the 
existing projection on the south. Fig. 142 shows the exterior 
aspect of the eastern part of the nave when seen from the 
south. The part that projects southwards is from the exterior 
quite worthy to be termed a transept, yet the archway leading 








to it from the nave is only 4 ft. 5 in. wide with a height of 
about 10 ft. 

This archway, Fig, 143, is of sufficient interest to repay a short 
digression. The wall in which it is pierced is 3 ft. thick, the 
voussoirs but not the jambs are formed of through-stones. Its 
Saxon origin may be argued from the inscription in Old English 

Fig. 142. — Breamore Church, Hants, from the south. 

which appears round the arch, and should be divided and 
translated as follows HER SPUTELAD SEO GECPX- 
DR^DNES DE, Here is-made-manifest the covenant to-thee.^ 
The meaning of the words is not clear. It has been suggested 
that they refer to the accomplishment of some vow, or again 
that they indicate the entrance to the baptistry, for which 
purpose the side space may have been used, though baptistries 
are generally at the western ends of churches. 

Passing on to other examples of the completely cruciform 
or undeveloped cruciform plans, we note that there is a fully 

^ The above is owed to the kindness of Mr. G. Gregory Smith. 



developed north transept at Stanton Lacy, Shropshire, but it 
has no corresponding transept to the south, and the access 
from the nave has been modernized. 

At Stow and Norton there are or were fully developed 
transepts, as large in section as the nave. At Dover and 
Worth the transepts are smaller though the churches may be 
considered completely cruciform. In respect of openings, at 
Stow the wide Saxon tower arches remain (see Fig. 146), but 



Fig. 143. — Archway into southern transept or chapel, Breamore, Hants. 

at Norton they have been enlarged ' by the removal of the 
inner order of voussoirs and those portions of the jambs which 
supported them,'^ a fact which implies a want of amplitude in 
the Saxon arches. At Dover again, the present transeptal 
arches of the twelfth century seem to represent an enlargement 
of the Saxon openings, while at Worth these openings are 
original. This leaves Worth the only Saxon cruciform church 
that Is in plan complete and untouched, and the example Is 
of sufficient importance to warrant a moment's pause. 

1 The Reliquary, Jan. 1894, p. 9. 


The general view, Fig. 145, taken from the south-east, and 
the plan. Fig. 144, will give an idea of the character of the 
edifice, the surroundings of which have inherited their wild 
sylvan beauty from the old forest of Andred that once covered 
this part of the country. The church is aisleless, cruciform 
and apsidal, but has no central tower, the tower seen in the 
view being a modern addition. The quoins are in long-and- 
short work and there is an abundant display of pilaster strips. 
A feature very pronounced in this building, that is not common 
elsewhere save in towers,^ is the horizontal string course which 
runs round the apsidal chancel as well as along the walls of the 
nave, though on the walls of the transepts, which are lower, it 
does not occur. The pilaster strips are bounded above by this 
string course where it appears, and above it come the windows, 
Li the nave these are of the double form, with the mid-wall shaft, 
illustrated in Fig. 39, ante, p. 92. There are original door- 
ways of characteristic Saxon type of nearly 14 ft. in extreme 
height by a width of 3 ft. 8 in., facing each other north and 
south towards the western end of the nave. The northern 
door is now blocked. 

The chancel arch at Worth is the finest of all that are 
extant. Its width is 14 ft. i in., its height 22 ft., and the rock- 
like massiveness of its huge ungainly imposts, and the large 
stones of the arch that take the whole thickness of the wall, 
are thoroughly Saxon. The arches into the transepts, in width 
8 ft. 7 in. (south) and 8 ft. 8 in. (north) by about 15 ft. in 
height, are plainer but equally imposing in their solidity. As the 
internal width of the transepts is 14 ft. 9 in. and 14 ft. 10 in. 
respectively the arches are of full proportionate size. 

We have now had before us a series of side chapels and 
transepts extending from the tiny lateral projection at St. Martin, 
Canterbury, entered merely through a doorway, to the im- 
posing transepts of Stow and the transeptal arches of Worth. 

^ It occurs at Bradford-on-Avon, Repton, and Wroxeter, and has left traces at 
Barholm, Lincolnshire. 










Breamore, and Deerhurst, where are transepts, but only door- 
ways into them, furnish significant intermediate links. 

(2) Nothing has yet been said about the central tower in its 
relation to the development of the plan. In idea the central 
tower, or some similar feature to emphasize the crossing, is an 
essential element of the developed cruciform plan, and it forms 
a conspicuous feature of our own later cruciform churches, as 
well as of those of Germany. As a fact however, in the 
evolution of the cruciform plan in Saxon architecture the 
central tower plays no decided part, and its non-appearance 
in the most perfect example of this plan at Worth is significant. 
Of churches which possess fully developed or embryo transepts, 
Dover, Norton, Stow, and Breamore still exhibit such towers, 
while their existence in Saxon times at Stanton Lacy, Repton, 
Deerhurst, and other examples, is problematical. If the Saxon 
central tower always involved for its support a thickening of the 
walls, its previous existence would in this way be detected. 
In the case of the axial towers of Dunham Magna (between 
nave and chancel) and Hooton Pagnell (at western end) there 
is this thickening, but it is not to be observed in the Saxon 
examples of Dover and Breamore, where the tower walls are 
no thicker than those of the nave. Hence this test of the 
previous existence of central towers in churches which now 
lack this feature, cannot yield sure results. The thickening 
may be held to prove the tower or the intention of one, but the 
absence of any such preparation in the walls for an additional 
superstructure does not preclude the actual existence of the 
central feature. In the case of Barton-on-Humber, where 
the tower was an original feature, the tower walls are the 
same thickness (2 ft. 6 in.) as those of the western adjunct 
and chancel. It was clearly not always considered necessary 
to thicken walls that had to be carried to the height of a 

The central tower may accordingly be regarded independ- 
ently of the cruciform plan, and when we take it in itself we 












' =§• 




see at once two distinct types of the feature. Going back for 
a moment to the axial tower as represented at Dunham Magna, 
we note that this is on the exterior flush with the walls of the 
nave out of which it rises without any break or feature. Only 
on the inside does it betray its existence in the plan by a 
thickening of the wall. On the east where the chancel is 
narrower than the tower the quoins of the latter descend inde- 
pendently to the ground. The same is the case at Dover, 
though here there is no thickening. At Breamore the tower 
not only asserts itself like the two last at its eastern quoins, but 
exhibits also what Dunham lacks, a continuation of the western 
tower quoins down to the ground in the form of courses of 
long-and-short work embedded in the wall and flush with the 
general surface of it (see Fig. 142, ante p. 234). A tower 
that is merely a growth out of the nave wall, like the three 
just mentioned, is independent of transepts, which may or may 
not exist without the tower being affected. Dunham Magna 
and Dover are almost alike so far as the tower is concerned, 
but only the latter has transepts. (See plans on pp. 225, 307.) 

On the other hand at Norton and Stow the central tower is 
quite distinct from the nave, and all four quoins of it descend 
clear to the ground in independence of any other structure.^ 

Fig. 146 indicates the plan of the Saxon central tower 
at Stow, with the nave, chancel, and transepts that are 
all narrower than the tower and abut against its walls 
leaving the corners free. This independence of the tower 
is emphasized also in the remarkable structure at Wootten 
Wawen, near Stratford-on-Avon, Warwickshire. A square 
Saxon tower stands here in the axial position between a nave, 
the north wall of which is old and may be contemporary 
with the tower but which has a later south aisle, and a later 
chancel with extensive south aisle. The tower is a good deal 
narrower than the nave, and narrower too than the present 

^This was noticed by Mr. C. C. Hodges in connection with Norton. 
The Reliquary, January, 1894, p. 9. 


chancel, but it possesses the remarkable feature that it has 
Saxon archways opening on each of its four sides as if it were 
intended to be a central tower with transepts to north and 
south. The dimensions of these arches are as follows — The 
western is 6 ft. lo in. in width by a height of a little over 14 
ft. ; the eastern 4 ft. 8 in. wide ; the northern and southern 



Fig. 146. — Sketch-plan of central tower at Stow, Lincolnshire, with parts 


each 4 ft. 2 in. There is no sign of contemporary side chapels 
or transepts. All these arches have plain square imposts 
with strip-work round the openings and long-and-short 
technique in the jambs. 

(3) If the reader will turn to the plan of the Saxon abbey 
church at Peterborough, given in Fig. 172, postea, p. 315, he 
will see a transeptal scheme strikingly different from those 
hitherto studied. Here are no transeptal arches and no pos- 
sibility even of a central tower, while the transepts are much 
ampler than the narrow presbytery, which excavations showed 


could not have extended much further to the east than the 
side walls now indicate. This plan, noticed postea p. 238, is 
apparently an example of the crux commissa or T shaped 
scheme, known in Early Christian days, and stands out of 
the line of development just followed. 

It is not the purpose here to enter into any analysis of 
these various schemes but only to indicate them as the data 
upon which must be based any discussion of the evolution of 
the cruciform plan. 

XI. Churches with both a Central and a 
Western Tower. 

The combination of the western with the central tower in 
the same building is a phenomenon of which account must be 
taken. It appears on a monumental scale in English Norman 
architecture at Ely, and the question naturally arises whether it 
is not an English form handed down from pre-Conquest days. 
There existed one conspicuous Saxon instance of it at the 
abbey of Ramsey in Huntingdonshire, dating about 970 a.d., 
of which a notice was given in the previous volume.^ The 
accounts contained in the Chronicle of Ra?nsey'^ of the building 
of the abbey church are instructive enough to merit quotation. 

The decision to erect a fine church upon this site in the 
Fens was arrived at in the summer of 968 and ' all through 
the following winter they are getting together all that the 
forethought of the masons demanded, whether in tools of iron 
or tools of wood, and everything else that seemed needful for 
the future building. At length v/hen the winter is past, the 
storehouses are thrown open, the most skilled workmen 
available are brought together, and the length and width of 
the church which is to be built are measured out. The 

^ Vol. I, p. 245 f, 

"^Chronicon Abbatiae Rameseiensis, Rolls Scries, No. 83, p. 38 f. 



foundations are dug deep on account of the marshy character 
of the site, and the earth is beaten with many strokes of the 
rammer to solidify it for the support of the weight. The 
labourers, inspired as much by the warmth of their pious 
devotion as by the desire for pay, are instant with their toil ; 
while some bring the stones others are mixing mortar, and a 
third party raises both stones and mortar aloft to the work by 
the aid of pulleys, and so with the help of the Lord the 
structure rises daily higher and higher,' 

' Two towers soared on high above the roofs, of which the 
lesser one was at the western end, on the front of the building, 
and offered from afar a stately spectacle to those entering the 
island ; while the greater one, in the centre of the cruciform 
structure, stood upon four pillars which were joined together 
pair and pair by arches thrown across from wing to wing to 
preserve the rigidity of the fabric. Compared with the old 
fashioned method of building which had before prevailed, it 
was a structure of no mean pretension.' The dedication of the 
church took place in November 974. 

There is an additional notice of the same operations in a 
Life of Bishop Oswald of Worcester, the patron and inspirer 
of the work. It is there recounted how the bishop inaugurated 
the work of the church with the sign of the cross which is the 
pledge of our salvation, and how he accordingly began to 
construct the buildings on the site after the pattern of a cross. 
There was on the east on the south and on the north an arm 
of the cross (the Latin word is ' porticus ' and the passage 
illustrates the wide usage of this term, see ante, p. 129 f.), and 
in the midst a tower which was sustained by these projecting parts. 
The church was then continued westward from the tower.-^ 

As Ramsey was a direct result of the extension to England 
of the Cluniac movement of monastic reform ^ it is worthy of 
note that one of the two continental monasteries through which 

'^Historians of the Church of Tork, Rolls Series, No. 71/1, p. 434. 
2 Vol. I, p. 228 f. 


this movement chiefly affected England, Blandinium by Ghent, 
possessed in the tenth century a church with a western tower.^ 

With regard to other Saxon examples, it is possible that 
Deerhurst, Fig, 140, ante, p. 235, possessed at one time a 
central tower as well as the existing one to the west, and 
Mr. Micklethwaite claims that the incorporation of the old 
Roman pharos at Dover in the scheme of the cruciform 
Saxon church that lies to the east of it gave Dover also a 
western as well as its central tower.^ 

There is some indication that the Saxon monastic church at 
Chichester, to which the South Saxon see was transferred 
from Selsey in 1075, possessed a central and a western tower. 
The evidence is an ancient seal attached to the M.S. Cott. 
xii. 80, in the British Museum, and numbered 1469 in the 
Catalogue of Seals. Its interpretation is not however devoid of 

XII. The Twin-Towered Fajade. 

This scheme, prefigured at Aachen and employed in later 

Carolingian times (ante, p. 53 f), became 

^ common, perhaps normal, north of the 

^'''"A^^^ Alps, for the greater churches of the 

^ /) /\\ eleventh and following centuries (ante, 

I^^J^fif I P- 5^)- W^ should naturally look for 

s]i^i;|*| / examples in Saxon England, and here 

.^■^ mpjj^ again we have the evidence of a seal on 

an Exeter deed of 1133, figured and dis- 

FiG. 147. — The Saxon j u /^ ^m- • u- r • r 7 

n .u J 1 .17 cussed by (j. UJiver m his Lives of the 

Cathedral at bxeter. ' J 

Bishops of Exeter. It shows two flanking 
towers and a central pavilion, and seems to represent the 
monastic church founded by ylilthelstan and restored by Cnut 
about 1020 A.D. See Fig. 147. 

^Van de Putte, Annates Blandiniensesy Ghent, 1842, ad ann. 979. 
"^ Archaeologicat Journal, liii, 327. 


XIII. The Triple-Apsed Plan. 

The termination of a church to the east with three parallel 
apses of which the middle one projects beyond the two lateral 
apses may depend on two different schemes of plan. The 
three apses may be the finish of a chancel and choir aisles, or 
they may result from a cruciform scheme when eastern apses 
are thrown out from the walls of the transept so as to flank the 
apse of an aisleless presbytery. Both these schemes are 
fairly early and widely diffused in the West. 

In Saxon architecture the form may be indicated at the 
cathedral of Oxford, once the church of St. Frideswide's 
nunnery. Here in the eastern wall of the present north aisle 
of the choir are three blocked archways that were discovered 
some years ago by Mr. Park Harrison, who also found traces 
of corresponding apses. It is not certain what they indicate, 
but they may point to an eastern termination for the earlier 
Saxon church on the spot similar to one of those familiar 
on the Continent ; that is, either a plan with three parallel aisles 
or with a transept with apses in its eastern wall, an alternative 
favoured by Mr. Micklethwaite.^ 

XIV. Aisled Churches. 

Only four existing Saxon churches can be proved to have 
possessed side aisles, though this has been conjectured with 
more or less probability about one or two others. On the 
other hand we can infer from literary sources that the larger 
and more important churches, which have in nearly every case 
perished, were laid out with these additions. This we know 
to have been the case with the Saxon cathedrals at Canterbury- 

1 Archaeological Journal, liii, 333. 

2 Willis, Canterbury Cathedral, p. 10. 



and Winchester,^ with the church at York described by Alcuin,^ 
with Wilfrid's churches at Ripon ^ and Hexham,* while the 
mention of a column in the interior of the abbey church at 
Ramsey * may be taken as evidence in this case also. 

The four extant examples are Brixworth, Northants ; 
Reculver, Kent ; Lydd, Kent, and Wing, Bucks. They are 
all pillar basilicas, in which the division between the nave and 
aisle is not formed as in the Italian churches and those of 
Africa by rows of columns, but by portions of the side wall 
left standing between the arched openings. Lydd, which is 
the simplest example, may be noticed first. The parish church 
of Lydd is one of the fine 
Gothic structures which 
make the district of the 
Romney marshes a favourite 
haunt of the ecclesiologist. 
A year or two ago it was 
discovered that the north- 
west corner of the north 
aisle of this well known 
church consists of part of the north and west walls of an early 
basilican oratory. Some of the arches and piers of the north 
arcade of its nave are visible in the present wall (Fig. 148). The 
span of the arches was about 4 ft., the width of the piers about 
3 ft. 4 in. An arched opening about 7 ft. wide can be traced in 
the west wall, now the end wall of the north aisle of the mediaeval 
church. This, it has been suggested by Mr. Micklethwaite, 
who was the first to call attention to the remains,^ may be the 
arch of triumph opening into an original western apse. This 
would furnish an interesting parallel to Silchester, to which 

^Willis, in the 'Winchester' volume of the Archaeological Institute, p. 15. 
^Historians of the Church of York, Rolls Series, No. 71/1, p. 394. 
2 ibid. p. 24. '♦ibid. p. 33. 

^ Chron. Abb. Ratnes., Rolls Series, No. 83, p. 104. 
^ Archaeolo^cal "Journal, lv. 

Fig. 148. — Saxon remains at the north- 
west corner of Lydd Church, Kent. 


perhaps as we shall presently see should be added Ripon and 
the earliest Canterbury. A round-headed opening, evidently 
double-splayed, exists in the old clearstory at A', of which 
A gives the plan and B the section. The plan shows the 
remains with details. C is the later tower. 

Brixworth, Reculver, and Wing all present features of 
special interest and importance and to these due attention must 
be paid. Brixworth is one of the largest and one of the most 
instructive of all our pre-Conquest monuments, but it has not 
come down to us in its original condition. On the one hand 
it has lost its side aisles though their previous existence is an 
obvious fact, but on the other it has received additions at one 
or more than one period later than that of its origin but still 
within Saxon limits. From the point of view of chronology 
this fact makes it of the highest value. Reculver, which had 
come down almost to our own time practically perfect, was 
ruthlessly destroyed at the beginning of the last century by 
one of the most shameless acts of vandalism ever perpetrated. 
Some details of great interest have however survived. Wing 
we are fortunate in possessing practically entire, with its nave, 
side aisles, and apse complete and in regular use, while beneath 
the presbytery exists a roomy crypt. It is not however so old 
a church by at least two centuries as the pair last mentioned. 

The general appearance of Brixworth church as seen from 
the north-west can be judged from the view. Fig. 149, while 
in Fig. 151 is given the plan. It is a large somewhat gaunt 
structure, the plain square rudely constructed fabric of which 
is crowned by an elegant spire of the 14th century. The use 
of Roman materials is obvious at a glance. All the openings 
are turned in Roman bricks, which are also employed here and 
there in the rubble walling more especially at the corners. A 
few courses in the rubble work are laid herring-bone fashion. 
A little observation of the manner in which these bricks are 
employed will show that they were certainly not placed by 
Roman hands. Generally speaking in the larger arches there 

1^^ 11 






are two rows of voussoirs, one outside the other, separated by 

flat courses of bricks concentric with the curve of the arch. 

In the two rows of voussoirs the bricks are set edgeways, and 

should all point towards the centre from which the curve of the 

arch is struck. Since the bricks are even, the mortar joints 

should be wedge-shaped to secure the form of the arch. Here 

at Brixworth however the principle of the radiating joint in 

arch construction was evidently not understood by the builder, 

and the manner in which he started 

to turn his arches can be seen in 

Fig, 1 50, where the bricks that 

begin the arch are shown tilted up 

at a sharp angle and wedged in 

position by a pad of mortar. There 

was a Roman settlement of some 

kind at Brixworth as numerous finds 

have attested, and from Roman 

buildings these bricks must have 

come. It has been claimed for 

the structure itself that it is a 

Roman survival, but independently 

of its technique it is in form and 

character an Early Christian basilica 

and there is nothing about it that suggests a building turned 

from a secular to a sacred purpose. When we learn that the 

monks of Medeshamstede (Peterborough) established a 

monastic settlement here about the year 680^ we have a date 

which, as we shall see, is in accord with the characteristics 

of the earliest parts of the building. 

The elements of the structure consist in a square western 
tower with a half-round stair turret projecting from its western 
face, a nave, a prolongation of the nave by a presbyterial space, 
and beyond this to the east, not seen in the drawing, an 
externally polygonal apse. Along the side of the nave on the 

^ Hugh White, in Sparke's Historiae Anglicae Script. Var. 11, 8. 

Fig. 150. — Springing of arch 
at Brixworth. 


ground story will be noticed a series of large arched recesses in 
each of which there appears a window. These are in reality 
the original openings which gave access from the nave to a 
north aisle. This north aisle, to which corresponded one on 
the south, must be restored in thought with its sloping lean-to 
roof abutting on the wall of the nave on a line marked by the 
sloping set-off in the wall under the upper row of windows. 
These last are the windows of the original clearstory, and are of 
a type uncommon in this country. They resemble the windows 
of the Early Christian basilicas of Rome and Ravenna in their 
openness and ample dimensions. The aperture is wider in the 
interior than it is outside, but the splay is nothing approaching 
to that which is seen in the ordinary internally-splayed lights 
of Late Saxon and of Norman times, while the actual width of 
the external aperture, which measures about 3 ft, in the clear, 
is much greater than we generally find in our Saxon buildings.-^ 
In comparison with the normal widely-splayed openings they 
present the appearance of being cut straight through the wall 
as is the fashion of the classical aperture. Strictly speaking 
they are not so formed, but exhibit, like all the windows we 
have been dealing with, the mediaeval feature of the splay. 
The stair turret is lighted by small square-headed openings, 
into the outer aperture of which were fitted stone slabs pierced 
with narrow loops. The suggestion has been made that 
these were loopholes for archers defending the church against 
attack. They do not however command the accesses to 
the building, and would have been almost useless for such a 
purpose. The plan of one opening is given in Fig. 154, 
postea, p. 255. 

The tower is one of those that, like Barton-on-Humber, 
possessed doorways or at any rate arched openings on all its 
four faces. A detail not to be passed over can be seen at the 
bottom of the north-west corner of the tower — that nearest the 
eye in the view — and also at the corresponding south-west quoin. 
1 The windows at St, Peter-on-the-Wall are similar, ante, p. 117, 

Fig. 151. — Plan of Brixworth Church. 


Here the attachment of walls thinner than those of the tower 
running in the directions north and south can be plainly seen, 
the lowest courses being' in bond with the tower. These 
indications of buildings now destroyed at the western end of 
the church are important and must be taken in connection with 
similar marks seen elsewhere. Thus in the case of the two 
very early porch-towers at Monkwearmouth and Corbridge 
there are signs of similar structures, and these are more 
apparent still on the very late western tower at Netheravon in 
Wiltshire. The significance of these indications is a matter 
for further inquiry. It will be noticed in the view that the 
original nave arcade and the clearstory windows stop short 
before the eastern end of the nave is reached. The walls of 
the side aisles, the foundations of which have been laid bare, 
also stopped at the same point, and these are the external 
signs of a peculiar feature of the interior plan which will 
presently receive attention. 

The church is entered through a round-headed doorway at 
the western extremity of the south wall of the nave that is 
inserted into one of the old arcade openings to the side aisle. 
The interior view of the church looking eastward reveals a 
temple of imposing size with a length to the altar of more 
than a hundred feet and a width of over thirty. The space is 
however not unbroken, for at a distance of about sixty feet 
from the western end there is a cross wall now broken by a 
single wide pointed arch dating about the year 1400. Origin- 
ally, as was proved by excavations in 1841, there was here an 
arcade of three arches supported by two intermediate piers and 
by the piers which still exist as projections from the north 
and south walls. This screen cut off a space of, roughly 
speaking, thirty feet square before the arch leading into the 
apse. In the north wall, just on the eastern side of the 
projecting pier, there is a narrow doorway 3 ft. wide, now 
blocked, and this would have led either into the open, or into 
some sacristy or similar building at the eastern end of the 


north aisle which stopped at the level of the projecting piers. 
What there was in this part on the south cannot be known as 
there is here a later mediaeval south chapel, and this, with the 
exception of the roof, the upper part of the tower and the spire, 
is almost the only portion of the structure that is not original, or 
at any rate restored, Saxon work. Passing on still eastwards 
we come to the apsidal presbytery in connection with which 
there is more than one point of interest.^ The arch of triumph, 
to use basilican terminology, is 9 ft. 8 in. wide with a height 
of nearly 22 ft. above the floor in front of it, and gives access 
to a chancel with an apse bounded internally by a semicircle 
but on the exterior by five sides of a polygon. A straight 
piece of wall before the semicircle and the polygon begin gives 
additional depth to the presbytery which measures 19 ft. 2 in. 
from west to east by a width of 1 8 ft. At the external angles 
of the polygon there are buttresses 18 in. wide with a pro- 
jection of 6 in. that are neatly cut to the form required. Of 
the actual walling of the apse only a portion on the north is 
original the rest being a restoration. A buttress occurs how- 
ever in the original work, and included in this also there was a 
round-headed window, now blocked and invisible, the situation 
of which is seen on the plan. There is no sign that the apse 
ever was vaulted. The present roof is of plaster. 

Returning to the west of the arch of triumph we notice on 
either side of it a round-headed internally-splayed window, and 
beneath these on each side a very low blocked doorway. On 
the exterior it is seen that these doorways once opened on, or 
rather in a flight of steps, for these must have begun within the 
church, that led down to an ambulatory or passage round the 
outside of the apse the level of which is about 6 ft. below 
the floor of the church. As can be seen on the north side this 

1 There is a valuable paper ' On the Chancel of Brixworth Church ' by Sir 
Henry Drj'den, in Ass. Soc. Reports, 1890, from which some of the measure- 
ments and details given in the text and in Fig. 152 have been derived. The 
paper notices previous publications on the church. 



passage was originally vaulted over and formed below ground 
a covered way round the apse that corresponded with the 
semicircular passages round, but as a rule within^ the circuit of 
apses in Early Christian churches on the Continent. Such 
passages generally give access to a small chamber called a 
'confessio'^ excavated under the floor of the apse for the recep- 

FiG. 152. — Sections at eastern end of Brixworth. 

tion of a sarcophagus or relics, but investigations have not 
revealed the presence at Brixworth of any chamber of the kind. 
On the other hand in the exterior wall of this ambulatory 
which is partly original there are two arched recesses that seem 
intended for tombs, and correspond to the ' arcosolia ' of the 
Roman catacombs. The arrangement of the presbytery and 
ambulatory, with the different levels, can be seen in the outline 
sections in Fig. 152, where a shows a section in front of the 
arch into the chancel, and b a section through the north wall 

^ From ' confessor,' * martyr,' as the place where the remains of such a holy 
personage were bestowed. 



of the chancel and the ambulatory. The springing of the vault 
over the latter is clearly visible on the existing wall. 

Turning now our steps westwards we are confronted by the 
western tower and by certain problems which connect them- 
selves therewith. The view Fig. 153 (in which the modern pews 
and other fittings are ignored), shows the thoroughly basilican 
aspect of the interior, in which we must imagine the arches of 

Fig. 153. — Interior view of Brixworth, looking west. 

the ground story opening into side aisles. The western end 
must now engage our attention. We see there on the ground 
story a doorway of entrance of a moderate size less than 5 ft. in 
width, but in the wall above it there are the marks of the 
springing of another arch apparently of about the same span 
the crown of which, as seen on the western face of the wall, 
rose to about 20 ft. above the floor. It will be remembered that 
traces of a blocked opening of much the same kind and 
position and of considerable width has come to light in the 
west wall of the nave of St. Martin, Canterbury (ante, p. 121). 


It is conceivable though unlikely that we have to deal in each 
case with a single very lofty arch of entrance. The arch of the 
present western doorway at Brixworth presents however no 
signs of having been inserted. 

Higher up than the indications of this blocked archway 
comes a triple opening the arches of which are divided by 
baluster shafts of a form already illustrated in Fig. 114, ante, 
p. 198. This opening cannot have coexisted with the archway 
just below and must be of a comparatively later date, which is 
also indicated by the form of the baluster shafts. Up to this 
point we have not found at Brixworth any of the familiar 
Saxon details such as pilaster strips, long-and-short work, or 
balusters. The appearance of these shafts is a sign that 
additions have been made in Saxon times to the main fabric, 
and we are led to remind ourselves of the probable history of 
the church. Founded about 680 as the church of a monastic 
settlement, but at the same time no doubt a parish church, 
and indeed from its size and situation a sort of mother church 
for the whole district, it was partially destroyed and rendered 
for a time useless by the heathen Danes. 

At a later period, probably near the days of revival under 
king Edgar, the fabric was restored to use but only as a parish 
church, and with the sacrifice of the side aisles, without which 
the building was of ample size for its purpose. To this 
epoch will belong the western opening with the baluster shafts. 

How much, we may ask, was done at this time ? It is 
natural to think that the whole tower, or at any rate the middle 
part of it, may be of the same date as the triple opening. Let 
us study this question in the tower itself. 

Entering this we note on the ground story the four openings 
on the four sides of the square. Of these the western one is 
now a mere doorway 3 ft. 6 in. wide, situated to the south of 
the middle of the wall and giving access to the turret staircase, 
hut marks in the wall show that it was originally a wide and 
lofty archway central in the wall, 6 ft. 8 in. in span with a height 


of 12 ft. 5 in. — a monumental outer portal to the imposing 
interior. It is quite clear therefore that just as the triple 
opening is later than the original western wall, so the stair turret 
is posterior in date to the ground story of the tower. What 
was the arrangement of this ground story before the turret was 
built .'' This question connects itself with that of the purpose 
of the subsidiary structures which were attached as we have 
seen to the western corners of the tower, and must not be 
entered on here, but the suggestion cannot be put aside that 
the lower part of the tower was originally a porch with a wide 
western doorway, and that the walls were carried up later in a 
tower form. The ground story of the tower, at any rate, 
with its western arch and with the now blocked archway in the 
western wall of the church noticed above, appears contem- 
porary with the rest of the fabric. An examination of the 
lower part of the western wall of the nave, in connection with 
the parts of the tower that are on the same level, exhibits such 
marked similarity in material and technique that all these parts 
must be of the same epoch. It is clear that the lowest story 
of the tower was a western adjunct of some kind. The date 
68o is too early for us to think of a western tower, but this 
feature would agree very well with the time of restoration, 
when we must suppose the walls of the western adjunct carried 
up in the form of a tower, and the triple opening cut in the 
west wall of the church. 

The most natural explanation of the stair turret would be 
that it was made to provide easy access to the chamber in which 
was the triple opening, and we must accordingly assume it to 
be contemporary with the tower, though it has often been 
placed at a later date. One thing is clear, the doorway, about 
3 ft. wide, that now gives access from the turret to the 
chamber with the triple opening must have been made in con- 
nection with the stair, for like the similar doorway on the 
ground story it is not in the middle of the wall, while, as the 
plan of this stage of the tower, Fig. 1 54, will show, the tower and 


turret seem very much of a piece. It is greatly to be deplored 
that the original belfry openings have not been preserved in 
the upper stage of the tower, as the details here might have 
furnished a valuable criterion as to date. 

It is not necessary to describe the plan of Reculver or to enter 
on the question of the treatment of the building in modern 
times. Of both there is a full account in Mr. Romilly Allen's 
Monumental History of the 
British Church} It was a 
simpler structure than Brix- 
worth, a pillar basilica ter- 
minating with a semicircular 
apse, about 75 ft. in interior 
length, with a nave and an apse 
24 ft, in width. It is note- 
worthy that the original con- 
crete pavement is said to 
remain below ground over Fig. 154. — Brixworth Tower at level 
almost the whole of the °f ^"ple opening, 

interior, and this is faced with fine plaster made with red 
pounded tile and brought up to a polished surface.^ The same 
sort of material, obviously a survival of Roman technique, 
covers, it will be remembered, part of the walls of the nave of 
St. Martin, Canterbury. Opus signinum similar in appearance 
may be seen on the floors at Uriconium and other Roman sites. 
The chief interest of the church is the treatment of the feature 
which in the basilicas of the Romanized West generally is called 
the arch of triumph. This is the archway which gives access to 
the apse, and is normally a single span the full width of the 
apse, the semi-dome of which abuts against it. Here however 
in place of a single arch spanning the nave at its eastern limit 
where the presbytery begins, there was an arcade of three arches 
that was brought forward a few feet into the nave so as to 
screen off a space before the apse. The three arches rested in 

^London, 1889, p. 13 f. "^ Archaeologia Cantiana, xii, 248. 



the middle on two detached columns that are among the most 
interesting and at the same time puzzling monuments that 
have come down to us from Saxon times. The plan, Fig. 155, 
gives this part of the building. The Saxon work^ is shown in 
black and the dotted lines extending eastward beyond the apse 

Fig. 155. — Plan of Saxon work at the eastern end of Reculver Church, 
Kent. From plan by George Dowker, F.G.S. 

indicate the Early English chancel of considerable length that 
was substituted for the apse in later mediaeval times. The 
screen of three arches was allowed to stand while the apse itself 
was removed, and remained till the destruction of the church 
in the early years of the nineteenth century. The columns 

^ Identified as such by Mr. George Dowker, and published by him in Arch. 
Cant. loc. cit. Much of the walling now seen above ground on the site is later 
than Saxon times. 


which divided the openings, after some migrations, were re- 
covered through the agency of the late Dr. Sheppard of 
Canterbury and now stand in the garden to the north of the 
north transept of Canterbury cathedral. Till quite recently 
they were accepted as Roman work, and on the strength of 
this assumption Reculver, like some other Saxon churches, was 
supposed to be in part a Roman secular building turned at a 
later date to Christian purposes. On this theory in general 
a word will be said on a later page, but apart from this, there 
is no reason why Roman columns should not have been used 
in a church itself of Saxon date. The Roman origin of the 
columns has however been contested, and the leading English 
authority on Roman architectural remains, Mr. G. E. Fox, 
has brought forward very forcible arguments against a Roman 

A word about Saxon columns in general may here be intro- 
duced. The subject is one of some importance, in view of the 
theory that the use of large round piers of columnar shape in 
Anglo-Norman architecture is due to a traditional familiarity 
with this feature on the part of Saxon builders. As a matter of 
fact, notices of the employment of such columns in pre-Con- 
quest buildings, as well as actual examples, are singularly hard 
to find. Wilfrid used columns in the seventh century in his 
churches at Hexham and Ripon,^ and in both these cases, no 
doubt, he derived them from the Roman stations on the Wall 
and in Yorkshire. The church begun by Archbishop iElbert at 
York in the eighth century was also 'solidis sufFulta columnis,'^ 
but we hear nothing of columns in connection with the later 
work at Winchester in the tenth century, or at Canterbury.* 

^ In a letter to The Builder of Oct. 20, 1900, and in Archaeological Journal, 
LIU, 355. 

''-Historians of the Church of York, Rolls Series, No. 71/1, pp. 24, 33. 

'^ Historians, etc. loc. cit. p. 394. 

*Sec the 'Winchester' Volume of the Archaeological Institute, 1845, and 
Canterbury Cathedral, Lond. 1845, both by Professor Willis. 


Of columns proper actually in use as supports in Saxon 
structures there are the four in the crypt at Repton, which have 
twisted monolithic shafts, round discs for bases, like those of 
Egyptian columns, and square capitals rudely chamfered off to 
fit the top of the shafts. Besides these there are preserved in 
the porch of Repton church two shafts built up of drums with 
similar capitals, that may have stood originally in the openings 
between nave and transepts or transeptal chapels. The only other 
examples that can be referred to are these two much discussed 
columns from Reculver church. They must be either Roman 
or Early Saxon. Mr. G. E. Fox considers that they are 
probably Saxon imitations of Roman work ; ^ but considering 
the limited technical capabilities of the Saxon builders, it is 
difficult to see how they could have executed such careful work 
on so large a scale, for the columns are more than i6 ft. high, 
and the shaft is everywhere within an inch of 7 ft. in circum- 
ference. The absence of tapering and entasis is quite abnormal 
for Roman shafts of this size and finish, though the ruder mono- 
liths already referred to in some churches in the north appear to 
be devoid of these refinements (ante, p. 8). It is true that no 
Roman parallel can be quoted to the form of the caps and 
bases, and that they both look barbarous beside the attic base of 
the Roman shaft to be seen at St. Pancras, Canterbury, in a 
position corresponding to that once occupied by these columns 
at Reculver. If we assume however that the capitals were 
intended ultimately to receive enrichments in gilded bronze, 
there is nothing unclassical about them or about the bases, 
which after all are not so utterly unlike in profile to some forms 
of the grecian ionic base. It is moreover just as hard to find 
for the columns Saxon as Roman affinities, and till something 
resembling them is found elsewhere, it will be best to reserve 
the question of their origin as still unsettled. The under- 
cutting of the lowest member of the neck-moulding should in 
this connection not be overlooked, as the only Saxon parallel 

^ Archaeological Journal, l 1 1 1 , 355. 



known to the writer is a moulding on the south door of the 
pre-Conquest tower at Barton-on-Humber. If these columns 
are to be reserved, we are left 
with the Repton examples 
as the sole monumental 
evidence for the supposed 
predilection of the Saxon 
builders for this feature. 
There is no Saxon columned 
basilica like the Mauritius- 
Kirche, near Hildesheim in 
Germany. (See Fig. 156.) 
The last of the existing 
aisled churches to be noticed 
is situated at Wing in 
Buckinghamshire. This has 
preserved its basilican plan, 
the eastern portion of which 
is shown in Fig. 160 postea, 
p. 268, but has suffered the 
loss of all its ancient open- 
ings, save a small blocked 
doorway at the end of the 
north aisle and a very in- 
teresting double window 
with a mid-wall shaft in 
the east wall of the nave 
above the presbytery arch, 
that has been already figured 
(Fig. 116 ante, p. 199). 
The external view of the 
church (Fig. 157) shows it 
a handsome modern-looking edifice without any Saxon character 
which would catch the eye. The later windows and the tower 
account for this. The polygonal apse is the most prominent 

Fig, 156. — Roman and Saxon columns. 

A. Roman monolithic shaft in Chollerton 

Church, Northumberland. 

B. Base of Roman column at St. Pancras, 


C. Column from Reculver Church. 

D. Moulding below capital. 

E. Base. 

F. Column at Repton Church, Derbyshire. 



feature in the view, and this it will be observed has round it a 
shallow arcading connecting a series of pilaster strips. Under- 
neath the whole area of this presbytery extends a crypt that will 
presently be noticed. Openings to it can be seen low down in 
the walls of the apse. 

Although this treatment of the types and features of Saxon 
churches is intended only to embrace existing monuments, 
some reference must here be made to an important aisled 

Fig. 158. — Willis's Plan of Saxon Cathedral at Canterbury, 
with some modifications. 

building, only known from descriptions, but known with 
sufficient clearness to entitle it to rank as a definite example. 
This is the Saxon cathedral church at Canterbury, of which 
Professor Willis evolved a plan from documentary evidence 
interpreted by him with his usual sagacity.^ The plan here 
offered (Fig. 158) is based upon that of Professor Willis, but 
there have been introduced some slight modifications. Thus 
the baptistry, or chapel of St. John, erected in the middle of the 
eighth century at the eastern end of the church, has been 
^ The Architectural H'tstovj of Canterbury Cathedral, Lond. 1845, p. 27. 


V ^ . >r.. 


'■\ ■ . 










made cruciform instead of octagonal. As the chapel had to 
serve for sepulture as well as for baptisms and other more 
secular purposes (Willis, p. 2), the cruciform scheme would 
be more convenient as supplying ' porticus,' as well as more 
in accordance with tradition.^ 

Again, the porch-towers, with the chapel of St. Gregory 
below the southern porch-tower, have been arranged more in 
accordance with known Saxon precedents. The plan of the 
crypt, indicated by dotted lines and the letters a, b, c (Fig. 158), 
has been drawn on a somewhat different scheme from that 
indicated in Willis's commentary (see postea, p. 267). 

The main features of the structure are distinctly indicated in 
the documents. It possessed ' alae ' or aisles, an apse at the 
east over the crypt, and various altars and flights of steps 
all clearly indicated in the original authorities which are 
printed by Willis in the opening pages of his work. There 
is no absolute indication that the church terminated towards 
the west in an apse, but the description makes this almost 

The question how the building came to assume the form 
indicated on the plan is one to which a little attention may 
be directed. Bede tells us that Augustine recovered in Canter- 
bury a church which he had learned was originally constructed 
by the labour of Roman believers, and constituted this as his 
episcopal seat.^ We have reason therefore to assume that 
the oldest part of the church was pre-Augustinian, though 
Augustine may have added to or altered it. The baptistry or 
chapel of St. John was erected by Archbishop Cuthbert, 740- 
758, while Archbishop Odo, 940-960, restored and heightened 
the edifice and modified the arrangement of the altars. 
Finally the church was ruined by a conflagration in 1067, 
and its remains seem to have been entirely cleared away by 

With the case of Silchester before us we may readily credit 
iVol. I, p. J 66. 2//X i, 33, 


the pre-Augustinian church with a western orientation, and the 

fact that the western end at Canterbury was raised some steps 

above the body of it, though there was no crypt in that part, 

may be explained on the analogy of the churches of North 

Africa where this raising of the altar end is not infrequent.^ 

At Canterbury the episcopal throne was against the western 

wall of the church, and the altar was away from the wall in 

front of it, an arrangement normal in the earliest Christian 

edifices. We may regard therefore this western part of the 

Canterbury church as a relic of Romano-British Christianity, 

while the basilican arrangement of nave and aisles may also be 

referred to the same source.^ The lateral adjuncts however 

with the south door, reminding us as they do of St. Pancras 

and of later Saxon plans, would appear to be additions made 

probably in the seventh century. At this time the adjuncts 

would at most be only porch-chapels, not towers. The south 

door, a later feature than the lateral chapel, may have been 

pierced in the southern wall of the chapel at a subsequent date. 

This suggestion for the early history of the church is borne 

out by the fact that the baptistry of the eighth century was 

erected at its eastern end. There is no reason for a baptistry 

to be placed toward any particular point of the compass, but 

there is a reason why the arrangements for this rite of admission 

to the Christian community should be located at the entrance 

end of churches. Hence we may infer that the eastern end of 

the church was at this time its place of entrance. Odo in the 

extensive works he carried out, which are said to have occupied 

three years, re-roofed the edifice and renewed and raised the 

upper part of the walls. Though we are not told that he 

altered the eastern end, we may conjecture that the apse in this 

part, with the crypt and confessio below it, was his work. 

^ Gsell, Les monuments antiques de P Algerie, n, 138. 

2 If Odo in the tenth century had added aisles to a single-celled church 
some mention would have been made of this in the notice of his work. At 
the time he took it in hand it is indicated as the largest church in the city. 


Such a crypt was impossible at the Romano-British epoch, 
though it might have been constructed in Augustine's time or 
in the eighth century. In either case the eastern would have 
ceased then to be the side of entrance, and the location there at 
the latter epoch of the baptistry would be unlikely. The apse 
and crypt are probably therefore later than the baptistry. 

The form of the crypt was evidently that of a curved 
passage, bb on the plan, following the line of the apse and 
communicating with a chamber or confessio, c, at the eastern 
limit. Such passages and chambers are found in Italy at an 
early date,^ but there is an example at Werden a.d. Ruhr in 
Rhenish Prussia that belongs to the ninth century. Not only 
the plan for the crypt, but the arrangement of the eastern 
end generally, might well have been derived by Odo from 
Germany, for such a duplication of the apse, by repeating it 
at the end of the building opposite to the original one, was, as 
we have already seen, a characteristic Austrasian feature, and 
in this connection it may be noted that the Early Saxon 
church at Abingdon had also an apse at each end.^ At the 
same time that Odo heightened the walls of the church he 
may have built towers on the walls of the already existing 
lateral porches. 

XV. Crypts. 

The last subject on the list is that of the Saxon Crypt. It 
is in one sense the most satisfactory to deal with, for the 
examples are few in number and comparatively well-known, 
while they present a series of types that are of value in connec- 
tion with the chronology of Saxon architecture in general. 
They are types that occur in numerous examples on the Con- 
tinent where it is possible to fix their approximate dates, and 

^ St. Peter, Rome, of doubtful date; Sant Apollinarc in Classe, Ravenna, 
siith or seventh century ; SS. Quattro Coronati, Rome, ninth century. 

* Chronkon Monast. de Abingdon, Rolls Series, No. 2/2, p. 272. 

264 . CRYPTS 

this naturally casts a welcome light on our own architectural 
problems. The known and existing Saxon crypts are those 
at Ripon and Hexham, at Brixworth, at Wing, at Repton, and 
at Sidbury in Devonshire, while to these should be added the 
crypt in the Saxon cathedral at Canterbury, which is of use as 
giving a form not fully represented in any other British 

The crypts at Ripon and Hexham, Fig. 159, are beyond 
doubt the work of Wilfrid, who was engaged upon the churches 
between the years 671 and 678, and we possess a contemporary 
notice of them from the pen of the bishop's own choirmaster 
Eddius. Eddius tells us about Hexham church that in its 
lower parts it contained certain chambers in the earth wrought 
of well-polished stones, and about Ripon that it was built of 
smoothed stone-work from the foundations in the earth up 
to the summit.- Structures in each case of almost exactly the 
same kind are seen at this day below ground on both of these 
sites, and except in Wilfrid's time there has never been any 
special connection between the places. When we find that 
these structures agree in style with the types current at that 
period in Europe generally, and in technique and material, at 
any rate at Hexham, with the special conditions of Wilfrid's 
work, there is no difficulty in deciding that we have before us 
what Eddius saw and described. There are indeed but few 
Early Christian monuments in Western Europe of which the 
date can be fixed with a certainty so absolute. 

Both at Hexham and Ripon the crypts are of a kind that 
have no necessary connection with a church, but occur in 

^ There was a crypt under the chancel of St. Olave, Chichester, a church 
that still keeps a Saxon-looking south doorway, but no signs of this crypt are 
now to be discerned. Other Saxon crypts probably exist and await discovery. 

^ — cujus profiinditatem in terra cum domibus mire politis lapidibus funda- 
tam. Fiia Wilfridi, c. xxii. Historians, etc. p. 33. 

— polito lapide a fundamcntis in terra usque ad summum aedificatam. 
ibid. c. xvi. Hist. p. 25. 





Fig. 159. — Crypts of Wilfrid at Ripon and Hexham. (The Hexham 
crypt is from a plan by Mr. C. C. Hodges.) 

A, A. Passages of access, now partly blocked. 

B, B'. Vestibules. 

C. Main chamber or confessio. 

D. (Ripon), Hagioscope ; (Hexham) stair from the west. 

E. Aperture in crown of vault (there was an opening of the kind 

also in B at Hexham). 
The spaces C, and B (Hexham), are covered with barrel 
vaults; B (Ripon) with a half barrel vault; B', B' 
(Hexham) with straight-sided vaults. The passages have 
flat stone roofs. 


simpler forms in Early Christian graveyards on the Continent 
(ante, p. 33). There is in each case an underground chamber, 
c, 12 or 13 ft. by 8 ft. covered with a barrel vault, with niches 
in the walls for the display of lamps. This was intended for 
the safe preservation and exhibition to the faithful of relics. 
Access was gained through antichambers bb', and these com- 
municated by narrow passages and flights of steps with the 
church above at aa. The antichamber b at Ripon is covered 
with a half-barrel vault in which there was an aperture at e. 
In the case of Ripon there is some indication of an opening, 
D, through which the faithful might gain a view into the crypt 
from the church above without actually making the descent, 
and if this be the case it would be clear evidence that the 
church was oriented towards the west. At Hexham d marks 
the beginning of a third flight of steps leading directly down 
towards the main antichamber b. In neither case can the 
connection of the crypt with the church be clearly made out. 

At Hexham the stones used in the construction are Roman, 
and many of them bear marks of Roman tooling and manipu- 
lation, while some exhibit carved ornaments and inscriptions. 
This is exactly what would be expected from the situation of 
Hexham close to important Roman stations. At Ripon, 
where Roman material was not readily available, the stone is 
expressly cut for the purpose and the technique here is sound 
and workmanlike. The walls and roofs throughout were in- 
tended to be plastered, and this material remains in parts in 
good preservation. It will be noticed that the passages of 
access are planned on rectangular and not curved lines. This 
does not prove, though it certainly suggests, that the churches 
above had rectangular and not apsidal terminations. Rect- 
angular crypts under apsidal presbyteries are not unknown. 
It should be noted that the southern passage of access at 
Ripon is only in its eastern part original. The letters a, a 
indicate where more modern work has at two periods been 


Equally early in design is the form of crypt already illus- 
trated at Canterbury, Fig. 158 ante, p, 260, and at Brixworth, 
Fig. 150 ante, p. 246. At Canterbury the passage followed 
the inner sweep of the apse and is described in the words 
via una quam curvatura cryptae ad occidentem vergentem con- 
cipiebat;^ at Brixworth it ran round outside the wall of the apse 
and was covered by a vault that abutted on this a little above 
the level of the ground. At Canterbury there was a chamber 
or confessio, c, to which this passage gave access, but no such 
chamber has been found at Brixworth. We know however 
that crypts were sometimes made under existing churches, and 
this was necessarily the case when a church had been erected 
in Early Christian times before the cultus of relics came into 
vogue. In our own country at Winchester in the tenth century 
crypts on a somewhat extensive scale were apparently formed 
under a church already built.^ At Brixworth the ambulatory 
seems contemporary with the presbytery itself, but it may have 
been intended to excavate the confessio afterwards. The 
ambulatory at Brixworth is entered from the two ends. At 
Canterbury to judge from the description the two ends were 
joined by a straight passage, b'. Fig. 158, forming the chord 
of the arc, but the position of the stairs or stair of access is not 
indicated. Their location at aa, Fig. 158, is conjectural but 
in accordance with precedent. 

The history of the crypt in general in European architecture 
exhibits a gradual opening out of originally confined spaces. 
The single vaulted chamber, a copy of a familiar form of the 
pagan Roman tomb, comes first, and narrow doorways or 
windows for inspection are the only openings in its walls. 
Later on the inner space grows larger to accommodate 
an increasing store of relics, and the wall is broken up 
into a series of piers between arched openings giving on the 
ambulatory. This stage of development is well represented 

^ Willis, p. 10, note. 

* Willis, in the 'Winchester' volume of the Arch. Inst. 1845, p. 13. 



in the noble Early Romanesque crypt at Montmajour near 

In our own country the crypt at Wing, Bucks, presents the 
same appearance though the technique is ruder. The plan is 
given in Fig. 1 60, where it will be seen that the voids, with the 

Fig. 160. — Eastern part of Church at Wing, showing the crypt in its 
relation to the structures above. 

necessary piers for the support of the floor above, fill the whole 
space within the walls of the apse. There is an ambulatory, 
the access to which from the upper church was by stairs at a 
and a', the opening for that at a being still to be seen. Fur- 
thermore towards the central chamber of the crypt, c, the old 
confessio, there opened a window of inspection, or hagioscope, 
D, visible now in the interior of the chamber, c. Though the 
church has been much modernized, the old arrangements can 



easily be made out and the levels offer no difficulty. The top 
of this western opening, or hagioscope, in the crypt is about 
level with the present floor of the nave and a little to the east 
of the chancel arch, so that access to it must have been gained 
by steps down from the nave-level in the centre of the flight 
which ascended on each side to the presbytery. This arrange- 
ment is indicated on the plan, which represents in parts the 
result of investigations made when the church was under 
restoration, e.g. the shaded parts in the arcade between nave 
and aisles show what in Saxon times was solid walline^. ee are 
openings that communicated with external tombs, or arcosolia, 
now destroyed. 

A further stage in the development of the crypt is reached 
when the heavy piers disappear and the necessary support to the 
roof is given by columns. This is 
what the Germans call the ' Hallen- 
krypta ' and it is the characteristic 
form in advanced Romanesque archi- 
tecture. The crypt under the altar end 
at Ste. Trinite, Caen, of the middle 
of the eleventh century, is a central 
example. Lastingham ; St. Peter, 
Oxford ; the cathedrals of Canterbury, 
Rochester, Worcester, Gloucester, 
Winchester, present us with Anglo-Norman specimens. It is 
not asserted here that the columned crypt is never found at 
an earlier date than is indicated in these examples. The crypt 
at Jouarre in France, which is of this kind, is ascribed by M. 
Enlart to the seventh century, and there are Early Romanesque 
instances in Germany. At the same time its place in architec- 
tural development is comparatively late, and for any example 
of the form in connection with Saxon structures an advanced 
date may be assumed. The only Saxon Hallenkrypta occurs 
at Repton, under a square-ended presbytery the external 
pilaster strips of which betoken a Late Saxon origin. The crypt, 

Fig. 161. — Plan of crypt 
at Repton. 


square in plan, Fig. i6i, is covered with rudely constructed 
vaults, partly groined, pardy segmental, with transverse arches 
resting on four central shafts and on projecting wall piers. 
The columns were noticed ante, p. 258. Passages of 
approach, one of which is still in use, opened from the two 
sides of the presbytery aa, and there existed as at Wing, a 
central hagioscope d. Noteworthy features, which also occur at 
Wing, are openings in the external walls of the crypts to north, 
south and east communicating no doubt with tombs situated on 
the exterior. There are indications of such tombs also in the 
outer wall of the ambulatory at Brixworth, Fig. 1 50 ante, p. 246. 

The architectural history of the crypt at Repton, as of the 
church above it, is obscure. There are obviously two periods 
of work in the upper parts of the crypt, but both appear to be 
late and not to be separated by any long distance of time ; on 
the other hand the masonry lining the lower walls of the crypt, 
at the eastern end, which is of large well-fitted stones, has 
been claimed as a possible relic of the pre-Danish monastic 
church (see vol. i, pp. 223, 227). ^ 

Finally at Sidbury in Devonshire there has recently come to 
light a small crypt of yet another form. Some details of this as 
well as its relation to the work above it betoken a Saxon origin. 

1 Repton forms the subject of a monograph by F. C. Hipkins, F.S.A., 
Repton, 1899. ^^- J- ^- Irvine gave much attention to the church and 
valuable drawings and plans of it are in the possession of the Society of 
Antiquaries of Scotland. Papers on it from his pen appeared in the Derby- 
shire Archaeological Society's y^ourna/, vol. v., and in xk^e Journal of the Archaeo- 
logical Association for 1894. Mr. Irvine doubted if the latest work in the 
crypt, including the columns and vault, was Saxon. An examination however 
of the external masonry of the eastern portions of the church, so far as they 
are preserved, indicates that they are all of the Saxon period. The north-eastern 
quoin of the nave (or of the tower if such existed) is treated just like the north- 
eastern quoin of the chancel, and the horizontal string course on the chancel 
can be traced round the former quoin and on to the north transeptal chapel. 
A Late Saxon date will suit both the columns and vault of the crypt and the 
rest of the walling better than a Norman date. Into other points of interest 
about the church and crypt there is no space to enter. 



JjN O R M A N 




162. — Crypt 


at Sidbury, 

It consists in a square chamber measuring about 10 ft. on each 

side, with a single stair of access towards the nave. This is not 

in the centre of the west side but to the north of the centre, so 

as to leave space to the south for a 

corresponding set of steps up from 

the level of the nave to the chancel. 

The floor of this which formed the 

roof of the crypt must have been 

of wood, for there are no signs 

of a vault. No niches or other 

features have been discovered 

(Fig. 162). The square-ended 

chancel above is Norman, and the 

crypt had apparently been filled in when the Norman walls 

were built, so that this fact, coupled with clear signs of 

Saxon technique in the jamb of the opening from the 

crypt to the stair, is good evidence of date. 

As regards type, we have here a form showing a still further 
advance in the direction of openness than even the columned 
crypt. This latter is as a rule only approached by two narrow 
passages, but in the form before us the crypt is open in front to 
the church and is directly accessible by comparatively broad 
stairways. In its monumental development the type is familiar 
in such examples as those at San Miniato, Florewce, and San 
Zeno, Verona ; and it is interesting to meet with the same type 
on a minute scale in an English country church. 



In the foregoing chapters there has been no attempt to offer 
a connected history of Saxon ecclesiastical architecture. The 
aim has been (i) to lay before the reader a collection of facts, 
grouped on some intelligible principle, that would serve as a 
groundwork for historical treatment, and (2) to offer 
hypotheses on which a consistent theory of this architectural 
phase might ultimately be constructed. The object before us 
now is to sketch the history and theory of Saxon building 
on the basis of the facts now ascertained and with the aid 
of the hypotheses already placed before the reader. It is 
desirable if possible to establish a chronology of the whole 
architectural period, and at the same time to fix the position 
of our pre-Conquest work in the general development of 
European architecture. 

The available chronological tests are those based on 
general appearance, technique, proportions, plan, and details. 
Of these the last test is in practice the most satisfactory. 
Inferences drawn from general aspect, or from the apparent 
place of a monument in an assumed scheme of development, 
are seldom really so cogent as the more direct testimony of 
some definite feature or piece of characteristic detail. It is 
proposed therefore in the first place to base this attempt at 


chronology on details, and then to test the resultant scheme 
by the other criteria just enumerated. 

As a starting point we may take the fact that at the 
present time most of those who have specially studied 
Saxon monuments are agreed in assigning an early date, 
in general terms the seventh century, to the following 
examples, St. Martin and St. Pancras, Canterbury (ante, 
pp. 119 ff.) ; Rochester (p. 119); Lyminge (p. 118); 
Reculver (p. 255); Brixworth (p. 246 f.) ; the crypts at 
Ripon and Hexham (p. 264 f.) ; Escomb (p. 1 10 f.) ; Monkwear- 
mouth (p. 140 f.), and Jarrow, and to this list many would 
add Bradford-on-Avon (p. 132 f.), and some Peterborough 

(p. 314 f-)- 

The evidence for the early date of these remains stands 
somewhat as follows. They have this in common that, with 
the exception of Bradford-on-Avon, they are distinguished 
by the absence of certain features which are common in 
Anglo-Saxon churches generally. We do not find in them 
long-and-short quoins, double windows with mid-wall shafts, 
double-splayed lights, pilaster strips, strip-work surrounding 
openings, plinths, nor, we may add, internally-splayed loops of 
a tall narrow form.^ If these features can be proved to be 
late, the absence of them may be taken as furnishing a 
presumption of comparative antiquity ; while on the other 
hand, if the buildings that lack these features can be shown 
on independent evidence to be early, this fact is evidence 
that the features themselves are comparatively late. It is 
of course no use to argue in a ring, and to endeavour to 
prove X early by the absence from it of Y, and Y late 
because it does not appear in X. We must seek for some 

^ Internally-splayed lights occur both in the earlier and in the later Saxon 
periods. In the earlier they are comparatively wide in comparison to tiicir 
height, e.g. west wall at Monkwearmouth, in the later they become tall and 
narrow, as at West Hampnett, Sussex, and approach a shape that occurs in 
Early Norman work, as at Darenth, Kent. (See Fig 41 ante, p. 93.) 
n S 


independent evidence of the comparative dates of the two, 
or our circle will be a vicious one. 

Fully to state and discuss this evidence would of course 
occupy far too much space, but a good portion of it has 
been already adduced in the notices of the buildings referred 
to. Only a word summarizing the evidence in each case 
can here be given, but the reader is assured that, if we put 
Bradford-on-Avon apart, and make a reservation in the 
case of the high walls of the nave at Monkwearmouth, none 
of the examples that are left present in their original work 
any feature that in the writer's judgment involves a later 
date than the seventh or eighth century. 

The crypt at Hexham, and by implication that at Ripon 
also, is proved by contemporary evidence to be of Wilfrid's 
time, or about 675. Roman stones of the same kind as 
are employed in the Hexham crypt are also used at Escomb. 
The character of the work at Monkwearmouth proves it to 
be prior to the Danish ravage of 867 (ante, p. 148 f.), while 
at Jarrow the western end of what is now the chancel of 
the modernized church agrees in the form of its oldest 
openings with Monkwearmouth. Turning to the south, at 
Rochester we find in the position of the earliest remains and 
their relation to the later structures a strong presumption 
of the date, about 606, suggested for them. The buildings 
at Lyminge, Reculver, and Brixworth are on sites where we 
hear of monastic establishments of the seventh century.^ 

At the two former places the remains show nothing incon- 
sistent with a corresponding date, while at Brixworth we have 
positive evidence of high antiquity (i) in the changes the 
building underwent in later Saxon times, (2) in the form of the 
original clearstory windows, which are of a type reminiscent of 

^ For Lyminge, c. 635, see Thos. of Elmham, //«/. Mon. S. Jug. Cant. Rolls 
Series, No. 8, p. 176. For Reculver, A.S. Chronicle, ad ann. 669; cf. Bede, 
Hist. Eccl. V, 8. For Brixworth, c. 680 Hugh White in Sparke, Hist. Angl. 
Scr. Far. Lond. 1727, 11, 8. 


Early Christian models (ante, p. 248). In the case of the 
Canterbury examples, the high antiquity of St. Martin as an 
ecclesiastical site is attested by Bede. The plan and technique 
of the oldest existing part would agree with a remote date, 
while the nave, which is clearly later, might well be pre- 
Danish. The tradition that St. Pancras was a very ancient 
church is not recorded at an earlier era than the fourteenth 
century (ante, p. 125), but the materials and technique of the 
building accord with the tradition, though the plan undoubtedly 
exhibits an advance beyond the simple Early Christian scheme. 

So much for the independent evidence of the early date of 
this group of monuments, evidence that derives a substantial 
part of its value from the undoubted authenticity of Wilfrid's 
crypts at Ripon and Hexham. We must now turn to the 
question whether we can fix to a comparatively late date the 
special Saxon features which do not appear in this group but 
are elsewhere so common. 

It is not necessary to add much in support of the opinion 
previously advanced (ante, p. 45 f.) that these features were 
introduced about the tenth century at the epoch when most 
of them were coming into general use in post-Carolingian 
Germany. The evidence in favour of this view seems so 
convincing that it is doubtful if in its general lines it will be 
seriously controverted. Will any one now maintain the theory 
that the Saxon pilaster-strips are copied from half-timber work, 
and not rather connected with the German Lisenen ? Or that 
Saxon towers, more than eighty per cent, of which are western 
towers, are derived from Italy where the western tower is 
almost unknown ^ Or that the windows with mid-wall shafts 
were fetched by a long journey from Italy when we could 
have found them, and found them too in western towers, just 
across the North Sea .'' Are we to claim double-splayed windows 
as our native invention, or credit them to Italy or Gaul where 
they are hardly found, when we know they were in abundant 
use in post-Carolingian Austrasia, and were there employed 



just as they are employed in England in constant association 
with the other features just reviewed ? There have been 
already indicated some of the more striking signs of affinity, 
such as the top of the tower at Sompting (Fig. 90 ante, pp. 
163, 200) and one of the caps in its openings (p. 202), and 
it may be further remarked that many of the cubical caps em- 
ployed in our bell-towers of the ' Lincolnshire ' type can be 
matched pretty exactly in the numerous Romanesque towers 


Fig. 163. — Pilaster strips and arcading. 
A. At Gernrode. B. At Earls Barton. 

of Germany. One of our most curious shapes, that in the 
western belfry opening at Glentworth, occurs at Gernrode and 
at Gelnhausen. The Scartho cap shown in Fig. 99, No. ix, 
is found, with the leaves rather more freely treated, in the 
oldest part of Werden a.d. Ruhr ; the eastern cap at Glentworth, 
ibid. No. VIII, appears at the Schloss Kirche at Quedlinburg. 
The arcading round the apse at Wing (Fig. 157 ante, p. 260) 
is mated by that on the chancel at Fischbeck in Westphalia.^ 
A comparison of A and B in Fig. 163 will show how like in 
character the Lisenen and arcading and string course on the 

^Lubke, die mittelalterliche Kunst in Westfalen, Leipzig, 1853, Taf. iii. 


north-western turret at Gernrode are to similar features on our 
own tower at Earls Barton.^ The question of comparative 
date and derivation may be passed over if we admit that our 
own use of the features and their employment on the other 
side of the German Ocean are parallel and, on the whole, 
belong to the same period. 

In the case of those features for which Austrasia provides 
no prototypes, the long-and-short quoins and the strip-work 
round openings, we have the facts that while they are not in 
evidence in the early group, they occur commonly in connec- 
tion with pilaster strips, double-splayed windows, and mid-wall 
shafts with which we must regard them as in the main con- 
temporary. An additional word or two on these particular 
features may be advisable. 

Though long-and-short quoins do not appear in the early 
group, yet the particular placing of stones involved occurs in 
openings, as at Escomb chancel arch and the western door of 
the porch at Monkwearmouth. When and through what 
stages the technique came to be applied to the quoining of 
walls we cannot at present determine, nor can we decide the 
relations in point of time between this system of making up 
corners and the employment in the same position of the large 
and massive quoin stones, as shown in Fig. 32 ante, p. 86. 
It is certain that these stones were in many cases Roman re- 
used, and where such blocks were available they would be 
employed at any epoch. There is no special evidence to fix 
the date ot St. Mildred, Canterbury, which furnished the 
example in Fig. 32, but Dover and Norton, where stones of 
the kind are used, are shown by their plans to be late. 

If long-and-short work appear to derive its origin from the 
early times, the other specially Saxon feature of strip- work 
round openings seems connected with the Austrasian feature 

^ See also the Frontispiece, which shows the straight-sided arches of the 
arcading. The same features appear on the nave wall at Gcddington, 


of the pilaster strip. It appears to be nothing more than the 
upright pilaster strip bent round the curve of the archivolt in 
the form of a hood mould, and if this be the case its chrono- 
logy will depend on that of the simple pilaster strip from 
which it is derived. The pilasters flanking openings start 
from corbel stones in the same peculiar fashion as the wall 
pilaster strips at Earls Barton and Stanton Lacy, Shropshire. 
(Compare Fig, 48 ante, p. 97, with Fig. 104, p. 187.) 

We thus obtain a useful line of demarcation between late 
and early Saxon buildings. Those in which appear the features 
just discussed are comparatively late ; while absence of these 
features combined with positive indications of early date suffice 
for the attribution of an example to the pre-Danish epoch. 
If we test this by the other criteria mentioned at the outset 
we shall find it confirmed rather than shaken. The criterion 
of general appearance might be taken to imply that a building 
is to be put at an early date because it appears rude and 
primitive in workmanship or is small in scale, and conversely 
that elaborate or workmanlike structures or those of large size 
are to be placed correspondingly late. In the case of Saxon 
England however, as noticed in the preface, there was no such 
normal social development as is assumed in such an application 
of the criterion. Some of the golden times of Saxon England, 
when the best and most ambitious work might be looked for, 
came at an early epoch. The direct evidence of more or less 
datable monuments is also against the criterion, for Wilfrid's 
crypts and the porch at Monkwearmouth are well wrought 
and cunningly enriched, and the early Brixworth is one of the 
largest churches produced during the whole period. 

The value of the criterion of proportion in ground plans 
may be judged from the evidence presented in the comparative 
chart, Fig. 30 ante, p. 84. If the reader will test the plans 
there given in outline by a rule placed diagonally from the 
right hand bottom corner common to all the plans, to the 
various upper corners on the left hand, he will see that 


(Jarrow Old Church excluded) the most elongated plans are 
those of Heysham Chapel and Monkwearmouth, while the 
two widest are Wareham and Rochester. Escomb and Coin 
Rogers come next for narrowness, and St. Pancras and St. 
Martin, Canterbury (the nave only), for width. Now both 
Monkwearmouth and Rochester, though so different in pro- 
portions, are both early, while the early St. Pancras and St. 
Martin almost exactly correspond with the late Wareham and 
Deerhurst Chapel. Coin Rogers is just as certainly late as 
Escomb is early, yet they are in proportions almost the same. 
Hence though elongated proportions may be used to mark 
off Saxon examples from Norman, they cannot be relied on 
for aid in subdividing the Saxon period. 

In the matter of proportions in elevation it has been already 
noticed (ante, p. 151) that height of walls is against an early 
date and suggests the Danish period, rather than the period 
of still unbroken Early Christian tradition. 

The criterion of type of plan as distinct from that of propor- 
tions is of more value as an indication of relative chronology. 

Of the types of plan noticed in the preceding survey (see 
ante, p. 100) some are simple and others complicated by the 
presence of towers or transepts. The simple plans include 
the plain rectangular oratory, the nave and chancel church, 
the apsidal oratory, the triapsidal, and the aisled or basilican 
scheme, and these may appear either early or late. The first, 
which we have found at an early date in Celtic lands, remains 
in use through the whole history of ecclesiastical architecture, 
and is this day a normal form for the private chapel of a 
mansion or institution. The example at Heysham (ante, 
p. 100 f.) may be comparatively early, and if we accept this 
it would involve an early date for the doorway at Somerford 
Keynes (ante, p. 102). The nave and chancel scheme, which 
in Ireland appears a natural development from the plain 
rectangle, may have been an Irish importation into Saxon 
England, though it must again be remarked that the Irish 


examples of the type have no great appearance of antiquity. 
It is at any rate a form that would be certain to appear in 
any Christian region where the Roman tradition of the apsidal 
ending was not in force, and even in Romanized lands it can 
hardly be considered a rarity.^ The absence of a tradition of 
vault construction would militate against the use of the apsidal 
plan, for this demands for its proper finish a semi-dome after 
the fashion that the French call a ' cul-de-four.' It is probable 
that the English backwardness in arch and vault construction 
(see ante, p. 127) has had as much to do with our national 
predilection for square-ended chancels as the ' Celtic tradition ' 
which is often invoked to explain this insular peculiarity. 

In any case, whether the rectangular presbytery be a 
natural growth or an importation from Ireland, it is no 
criterion of date or period as it occurs in the indubitably 
early Escomb and in the certainly late Repton and Boarhunt. 
The same may be said about the apsidal presbytery. The 
early Kentish group and Brixworth exhibit the apse, but so 
also do Worth and Wing, which are marked as late by 
their pilaster strips and mid-wall shafts. No chronological 
or geographical principle seems to be involved in the presence 
or absence of an apse. The ten examples enumerated in a 
note, ante, p. 118, are distributed pretty evenly over the 
country and are of various dates. 

The statistical relations between the apsidal and square- 
ended presbyteries are not easy to fix. The eastern ends 
of the vast majority of Saxon churches have been altered, 
and the subject is complicated by the fact that Early Norman 
apses were sometimes substituted for older Saxon western 
walls. At Bosham, for example (see Fig. 173 postea, p. 328), 

1 In the Romanesque period, M. Enlart remarks, ' les petites et moyennes 
eglises principalement en Picardie, Ile-de-France, Normandie, Champagne et 
Bourgogne, ont souvent un chevet rectangulaire ' {Manuel, p. 223), while 
Dehio and von Bezold note the predilection for flat eastern ends in the 
Romanesque of the Upper Rhine and Swabia {Kirchlkhe Bauku?tst, i, 208). 


the middle third of the extensive chancel shows by the 
herring-bone work in the walling that it represents an Early 
Norman extension of the smaller Saxon chancel. The latter 
in all probability was square-ended but the Norman chancel 
seems to have had an apse. The apses that are said to 
have existed at Dunham Magna, Norfolk ; and Corhampton, 
Hants, may be similarly explained. Taking the compara- 
tively few Saxon churches of which the eastern termination 
is assured we can count a score of square ends,^ to set 
against the ten apses already enumerated, but in all proba- 
bility the square-ended chancels out-numbered the apsidal 
ones many times over. 

The appearance of side aisles is no criterion of date, for, 
keeping only to existing monuments, we find basilican 
churches of the early period (Brixworth and Reculver), and 
also comparatively late (Wing, Buckinghamshire). It is how- 
ever a significant fact that so many early churches of ample 
proportions, such as St. Pancras, Canterbury ; and Rochester, 
are aisleless, whereas in the Romano-British period the 
narrower interior at Silchester was divided into nave and 
aisles (see the proportions compared in Fig. 30 ante, p. 84). 
The preference for the basilican form even for small churches 
is marked in North Africa as in all other centres of Early 
Christian art, whereas in Saxon England, though there only 
exist four aisled churches, there are nearly ninety about 
which we can be reasonably sure that in their original shape 
they were aisleless. 

Passing now from the simple plans to those which exhibit 
transepts or towers, we start with the side chapels, ' porticus,' 
and western porches, ' porticus ingressus,' which appear at 
St. Pancras and elsewhere. About these it has been already 

1 i.e. Barton-on-Humber ; Barrow ; Boarhunt ; Bradford-on-Avon ; Brea- 
more ; North Burcombe; Coin Rogers ; Daglingworth ; Dccrhunt Chapel; 
Dover; Escomb; Heysham Chapel; Kirk Hammerton ; Rcpton ; Sidbury ; 
Tichborne ; Wareham ; Weybourn ; Whitfield; Wittering. 



noted that the former seem to be the beginning of the 
development of the cruciform plan (ante, pp. 129 f., 226 f.) 
while the latter in some cases at any rate are the forerunners 
of the western tower. The early date of the ' porticus ' and 
' porticus ingressus ' are attested by their mention in Bede,^ 
so that the appearance of these features, say at St. Pancras, 
need excite no suspicion of the early date of buildings con- 
taining them. The cruciform plan and the central or western 

tower are on the other 
hand distinct indications 
of a comparatively ad- 
vanced date, and on this 
subject a word or two 
must be said. 

It will conduce to clear- 
ness if we glance first at 
the history of the cruci- 
form plan and of the tower 
in European architecture 

( I ) The reader's atten- 
tion has already been called 
to the special form of the 
sepulchral church of which Constantine's Apostles church at 
Byzantium seems to have been the prototype. The scheme 
of this and of the churches formed upon its model was 
founded on the Greek cross with equal arms, but as a matter 
of practice the arm opposite the altar is often extended to 
somewhat greater length than the others. 

Though Constantine's church has perished, an Apostles 
church ' ad modum crucis,' perhaps a copy of it, was erected 
at Milan at the close of the fourth century by St. Ambrose, 
and the plan of it has survived in the later Milanese structure 
of San Nazaro Grande (Fig. 164). The famous tomb of 
^ e.g. H.E. ii, 3, and Historia Abbatum c. 20. 

Fig. 164.. — Plan of San Nazaro Grande, 
Milan, Greek Cross. (No scale.) 



Galla Placidia at Ravenna, of the fifth century, is an example 
on a smaller scale of the same form. Here too the western 
arm is the longest. The chapel in the archbishop's palace at 
Ravenna, ascribed to the fifth century, is an interesting early 
example of the plan used for a building that was not sepulchral 
in intent. In buildings of this type the most suitable feature 
to mark the crossing would be the dome or pavilion, and an 

Fig. 165. — Plan of early Church 
of St. Denis, showing T 
form. (No scale.) 

Fig. 166. — Altar end of Church on Plan 
of St. Gall, showing Latin Cross. 

internal dome that externally takes the appearance of a square 
tower is used in the Ravennate building of Galla Placidia just 

(2) It was noticed ante, page 1 6, that an approach to the cruci- 
form plan is made in some of the Early Christian basilicas of 
the city Rome. The same scheme, whether derived from 
Rome or independently developed, comes into use among the 
Merovingian Franks, who seem to have favoured T formed 
buildings in the shape of the crux commissa, or cross without 
a head.^ Of this form, according to Viollet-le-Duc, was the 
original church of St. Denis, erected by Dagobert about 628 a.d. 
(Fig. 165). 

^ Enlart, Manuel d'' Arch'eohgie Fravfaise, i, 122, mentions several sixth 
century Gallic churches with transepts. 



(3) The ' crux immissa,' or complete Latin cross, makes 
one of its first appearances in the middle ages in the scheme of 
the church on the Plan of St. Gall, dating about 820 (Fig. 
166), Here the cruciform plan is not complete, but as it 
were in process of formation ; while it is more pronounced at 
Hersfeld (Fig. 167), the plan of which seems to date from 

about the middle of the 
ninth century. Even 
then the scheme is con- 
fined to the eastern part 
of the Carolingian realm, 
for the searching investi- 
gations of Dehio and von 
Bezold in their Kirchlkhe 
Baukunst des Ahendlandes 
have failed to find any 
example in Gaul or Italy 
earlier than the eleventh 
century,^ when we come 
upon the pronounced 
example of the cathedral 
of Pisa, and from this time onward the cruciform plan becomes 
normal for the greater churches all over Europe. 

It is clear that this development, though it may be con- 
nected with the Merovingian crux commissa, or to go further 
back with the transeptal Roman basilicas, is quite independent 
of the Greek-cross type of church. In the latter the four arms 
of the cross, the nave, transepts, and choir, are of the same 
width, height, and general importance, but in the buildings in 
the just-indicated line of development the relations between 
these different parts are at first quite irregular and accidental, 
and it is only later on that they are made to agree. For 
instance Hersfeld has its transepts narrower than its nave, but 
a little later, as at Wiirzburg, the dimensions have become the 

1 Kirchl. Bauk. i, 161. 

Fig. 167. — Plan of Hersfeld, of the ninth 
century. (No scale.) 


same. The idea of the cruciform church was as it were flashed 
upon Europe at an early date in the complete and consistent 
plan of the Greek cross, but the idea was not taken up, and 
the real cruciform church of the Romanesque epoch was only 
arrived at independently after many experiments. 

Turning to the question of the general history of the tower 
in Christian architecture, we find as already indicated that the 
early Greek-cross plans involved the feature of a central 
pavilion to mark the crossing, and that this pavilion took 
externally the form of a tower. The Gallic church of St. 
Martin at Tours built by Perpetuus about 470 a.d., and the 
Merovingian cathedral at Nantes described by Venantius 
Fortunatus showed this feature.^ These pavilion-towers 
represent the earliest form of the tower in Christian archi- 
tecture. We have seen already however that this early 
Greek-cross plan is not in the direct line of development 
which ultimately produced the Latin-cross plan of later 
mediaeval times, and the same may be said of the central 
pavilion, for this is not the same thing as the later central 
tower over the intersection of the arms of the Latin cross. 

In Christian architecture generally the second form of the 
tower seems to be the detached tower, which appears at 
Ravenna and Rome a little later than the pavilion-tower, 
and at the same epoch (not later than the seventh century) 
the germ of the twin-towered facade appears in the fore- 
building flanked with stair-turrets at San Lorenzo, Milan ; 
San Vitale, Ravenna, or in the East at Tourmanin in Central 
Syria. At the minster of Aachen of 796 a.d. the fore- 
building has already the form of a western tower flanked 
with turrets (ante, p. 51), while in the almost contemporary 
Plan of St. Gall, the flanking turrets are emancipated and 
stand as independent detached towers on each side of the 
entrance end of the building. The single western tower 
asserts itself, in Austrasia at any rate, at a somewhat later 

^ Enlart, Manuel, p. 122 f. 



date, and Dehio and von Bezold consider that it is in 
general the latest form of the tower that Romanesque 
architecture produces.^ 

As a result of these general indications of date the following 
propositions may be regarded as established. 

1. A Greek-cross plan with central pavilion-tower might 
belong to the earliest Saxon period. 

2. A crux commissa plan, with the clear transverse space 
across the transepts, might be equally early. 

3. It does not follow that the Latin-cross plan familiar in 
later times goes back to an early date, as this seems to be the 
outcome of an independent course of development. 

4. The central towers of the Latin-cross churches, and the 
axial and western towers generally, need not be placed early 
because of the early appearance of the pavilion-tower. 

In our own country, as in Europe generally, we find early 
examples both of the Greek-cross and T plans, for at the be- 
ginning of the eighth century Wilfrid erected at Hexham a 
church on the former plan with a central pavilion-tower (see 
postea, p. 319) and the T shaped scheme of Peterborough 
(postea, p. 315) may be of the same epoch, but these facts do 
not carry with them early dates for our cruciform churches and 
towers generally. 

No pavilion-towers or detached turrets have come down to 
us from Saxon times, but there are a dozen axial and central 
towers and nearly eighty western ones. Viewed in themselves, 
without reference to the details they exhibit but in relation 
merely to the general place of the tower in architectural 
development, they would all have to be placed in post- 
Carolingian times, and could not be dated earlier than the 
ninth or tenth centuries. Most of these towers however as 
we have already seen, like the cruciform churches, possess 
the characteristic details which indicate an advanced period, 

^ Die jiingste Erscheinung erst ist der Einzelturm an der Westfassade. 
Kirchl. Bauk. i, 561. 



and these details entorce the other indications of date just 
given. The fact that the use of the double belfry opening, 
so characteristic of Saxon towers, is carried on into post- 
Conquest times (Hornby, Yorkshire ; Boothby Pagnell, Lin- 
colnshire) is another proof of the comparative lateness alike 
of the tower and of these special features. 

Fig. 168. — Masonry of Roman character 
at Stone-by-Faversham, Kent. 

By the aid of these indications of date we may now be able 
to distribute most of the existing monuments among the three 
periods already indicated 
(ante, p. 35). 

Out of the dozen monu- 
ments mentioned on page 
273 we have for reasons 
already given (ante, p. 73) 
to reject Bradford-on-Avon, 
though we accept Peter- 
borough, and we may ask 
whether there are any other 
churches beside the last 
named that we shall be justified in including in the early list. 
One such we find at Corbridge (ante, p. 151) where the great 
portal, the Roman stones, the western porch, the internally- 
splayed lights, agreeing as they do with Escomb and Monk- 
wearmouth, place the building in the pre-Danish period. 

Stone-by-Faversham,^ which has not been mentioned in the 
previous chapters, gives us an example of technique that is 
perhaps more in the antique style than any other piece of work 
in the country. The ruins, consisting only in the lower courses 
of the walls, lie in a field a few hundred yards north of the 
main road from Faversham to Sittingbourne a little beyond 
Ospringe, and are worthy a visit, if only for the fact that the 
chancel, a mediaeval extension of the Saxon presbytery, still 
keeps at its eastern end the massif of the old stone altar. The 
^ Sec Archaeologta Cantlana, ix, Ixxvii. 


oldest parts are the western ends of the chancel walls and on 
the south the adjacent quoin of the nave. A bit of the 
masonry in the angle between the chancel wall and the east 
wall of the nave on the south is shown in Fig. i68. It is 
composed of squared blocks of tufa and Kentish rag, of the 
type of those used in continental petit appareil, alternating in 
classical fashion with Roman tiles. There is no difficulty in 
accepting this as of early date. 

Britford has often been claimed as early, and the use of 
Roman bricks with the curious mortising of jamb and impost 
in the south opening (ante, p. 229) give colour to the pretension. 
The plan, with eastern transeptal chapels, is however late, and 
a comparatively advanced date seems forced upon us by the 
stripwork, the presence of which round the openings is indicated 
by traces that are unmistakable. No other example seems to 
offer any positive evidence of early date, that is of the first 
period c. 600-800 a.d., though there are at the same time about 
a score of others which have no late indications, and may 
possibly belong, if not to the early period, yet to the 
intermediate or Danish epoch (800-950) from the commence- 
ment of the invasions to the time of Edgar. Of these the 
most interesting, because best preserved, examples are Avebury, 
Wilts (ante, p. 172) ; Bishopstone, Sussex (p. 131) ; St. Mildred, 
Canterbury (p. 86) ; Heysham Chapel (p. 100) ; Sockburn, 
Durham ; St. Michael, St. Albans ;^ Bardsey, Yorks (p. 156) ; 
Lydd, Kent (p. 245). The rest are more fragmentary.^ 

On the other hand the late indications already discussed, 
whether in the form of characteristic Austrasian details or of 
advanced features of plan and elevation, enable us to group 
together about one hundred and thirty churches as belonging to 
the later period, or from about 950 to the Conquest. Double- 

^ St. Michael was originally built by Abbot Ulsinus probably about 950. 
See Gesta Abbatum Mon. S. A/banl, Rolls Series, No. 28/5, p. 22, 

2 They comprise Hart, Durham ; Hackness, Yorks ; Oxford Cathedral ; 
Somerford Keynes, Wilts ; Stoke d'Abernon, Surrey ; Wroxeter, Salop. 


splayed windows, pilaster strips, and mid-wall shafts account 
for about ninety examples, while most of the others exhibit a 
treatment of ground plans, of openings, or of details that are of 
distinctly Romanesque character. 

It need not be assumed that the western tower is in itself an 
infallible sign of the latest of the three periods. By far the 
greater number of the existing towers are doubtless subsequent 
to 950, but some may be earlier, and among these the tower at 
Deerhurst is one whose chronological position will have to be 
carefully studied. 

About another score of the monuments present as their sole 
indication of pre-Conquest date the characteristic Saxon 
long-and-short work in the quoins. This we have found to be 
native in origin and evolved from the upright and flat slabs 
used in early archways as at Escomb and Monkwearmouth. 
How soon it made its appearance we cannot tell, but it does 
not occur in the early group. On the other hand we find 
it used frequently in conjunction with features that are 
acknowledged to be late, and it must be regarded as a sign 
of either the midmost or the latest of the three periods of 
Saxon architecture. 

To sum up the above, we can place some fourteen or fi.fteen ^ 
examples in the first period, and one hundred and thirty in the 
third, which will leave some fifteen that may be provisionally 
located in the intermediate period, while the long-and-short 
work of the rest puts them either in this period or the third. 
In the enumeration of monuments given in the index list of 
Saxon churches at the end of the volume, the letters A, B, C, 
as signifying the periods, are appended to the names in the list, 
and convey a more definite indication of the writer's opinion in 
the case of individual examples. This opinion it must be 
understood is in many cases only given provisionally and under 
reservations explained in the note postea, p. 331 f. 

^ Fifteen, if we include St. Peter-on-the-Wall, the difficulties connected with 
which were indicated ante, p. 1 16. 



The internal chronology of the third architectural period is 
not easy to treat. The monuments which belong to it are as a 
whole marked ofF by the signs we have come to know from those 
of the earlier periods, but among the acknowledged members of 
the group there seem at present only imperfect means for 
distinguishing early, middle, and late. These three sub-periods, 
O, O, O, correspond respectively to the epoch of revival in the 
latter part of the tenth century ; to that of church restoration 
under Cnut at the close of the Danish wars in the early part of 
the eleventh century ; and finally to that of the church exten- 
sion which seems to have gone on actively in the time of 
Edward the Confessor. The question is how the work of 
these different epochs is to be distinguished. 

As a starting point we may take the fairly dateable Barton- 
on-Humber (ante, p. 208 f.). It exhibits pre-Conquest work of 
two dates, so that the earlier, comprising the whole Saxon fabric 
save the uppermost story of the tower, may be provisionally 
placed in the first sub-period. An examination of the building 
seems to favour this date. We find here most of the characteristic 
late features in fully developed forms, but we note the retention 
of the earlier baluster shafts. As there is a continental prototype 
for the plan dating about 900 a.d., we may reasonably ascribe the 
work to the latter part of the tenth century. This ascription 
carries with it a suggestion of an equally early date for the two 
other towers enriched with pilaster strips and carving, at Earls 
Barton and Barnack. The builder of the first of these 
employed the banded baluster shafts of old tradition (Fig. 115, 
ante, p. 199) and divided his belfry openings in an abnormal 
fashion (ante, p. 189), though on a scheme different from that 
used at Deerhurst (postea, p. 300). The elaborate enrichment 
at Barnack has nothing about it that is necessarily late. These 
three towers are so different from the plain unpretending 
structures of the ' Lincolnshire ' type, with their characteristic 
mid-wall shafts, that we are naturally disposed to locate them in 
distinct epochs, and as the 'Lincolnshire' towers, which over- 


lap into the Norman period, are mostly ot the last sub-period, 
we have an additional reason for putting the exceptional towers 
correspondingly early. ^ 

It must be observed that, however elaborate be the work on 
the exteriors or in the tower arches of these exceptional 
structures, we do not observe either angle and soffit shafts, or 
developed roll mouldings such as are common in the advanced 
Romanesque of every land. The appearance of these Roman- 
esque features we may regard as betokening a date near the 
middle of the eleventh century, while when they are absent 
the work may be carried back to the end of the tenth. The 
tower arches at Barnack and Cambridge have moulded imposts, 
pilaster strips carried round as hood moulds, and in the latter 
case sculptured animals, but the arches are in both cases cut 
straight through the walls and there is no recessing or 
membering of the arch, such as we find at Wittering (Fig. 
59, ante, p. 108) Sompting (Fig. 118, ante, p. 201) or Bosham 
(Fig. 174, postea, p. 329). Similarly, the tower, formerly 
chancel, arch at Barton-on-Humber is plainly cut (Fig. 130, 
ante, p. 215), while that at Broughton is recessed and supplied 
with angle and soffit shafts (Fig. 128). This indicates a distinc- 
tion of period between the buildings though they are so alike 
in plan. 

So far as features and details are concerned, these criteria of 
the retention of balusters on the one hand, and on the other 
the use of angle or soffit shafts and recessing seem the only one 
available for separating the monuments of the Edgar period from 
those of Edward the Confessor's day. Additional investigation 

^ Besides the conspicuous examples like Earls Barton and Barnack there arc 
other towers in different parts of the country that do not exhibit the 
characteristic belfry openings (though in some cases they may have formerly 
possessed them) and some of these may be early in the third period. The 
following are some examples : — Bedford (St. Peter) ; Brigstock ; Cavcrsfield ; 
Guildford; Hough-on-the-Hill; Skipwith ; Stevington ; Swanscombe. 

St. Benet, Cambridge, retains the baluster shafts in belfry openings. 


in the future may enable the chronology of this last main period 
to be more accurately fixed, but for the moment the subject cannot 
be carried further, and this tentative chronological survey may 
conclude with a list (i) of a few monuments of importance that 
are brought by their pilaster strips or double-splayed lights 
within the compass of this period, but at the same time show no 
features incompatible with the earlier part of it ; and (2) a list 
of these buildings marked as specially late by the criterion just 

(i) Some examples that may belong to the early part of the 
third period : — 

Arlington, Sussex ; Bradford-on-Avon ; Breamore, Hants ; 
Brigstock, Northamptonshire ; Britford, Wilts ; Corhampton, 
Hants ; St. Mary, Dover Castle ; Repton ; Stanton Lacy, 
Shropshire ; Stow, Lincolnshire (lower part of transepts) ; 
Whitfield, Kent ; Wing, Bucks. 

A brief note on Stow may be permitted.^ To the place is 
attached a lordly tradition of early origin and episcopal rank, 
but we really know nothing about it till near the year 1040, 
when the then bishop of Dorchester, with the bountiful aid of 
Leofric and Godiva, set up there a religious establishment 
apparently of secular canons. We also hear of it about fifty 
years later, when Remigius the Norman bishop of Lincoln 
states in a charter that he has decided to renovate the place 
which was in a state of decay through the lapse of time and the 
neglect of those in charge.^ The present nave is generally 
reputed to be his work, while the fine vaulted chanceP is 
attributed to bishop Alexander in the first half of the twelfth 
century. The Saxon work is confined to the transepts 
and the remains of the central tower with its arches, and as 

1 Cf. Rev. G, Atkinson, on Stow Church, in Ass. Soc. Reports, 11, 315, 
and Gentletnan's Magazine, 1 863/1, 755. Also Dugdale, Mon. iii, s.v. Eynsham. 

2 Ecclesiam ... in loco qui vulgo dicitur Stowa, quondam prolixo temporis 
spatio praesidentium incur iadesola tarn, reformaredecerno. Dugdale, Mow.iii, 14. 

2 The present vault is a reconstruction. 


regards the dates of the different parts of this work, it is clear 
that the moulded archivolt of the tower arch from the nave 
cannot be earlier than the date c. 1040, and might be later. 
The piers of the arch may be earlier than 1040, but, from 
their pilasters, must be of the third period. The transepts 
however, in their lower portions, have been claimed as relics 
of a much earlier church, and so far as their plan is concerned 
they might conceivably be the relics of a crux commissa church 
like Peterborough. The internal length of the transepts is 
82 ft. as against the 92 ft. at Peterborough. Unfortunately 
however for this attractive theory, the transepts rest upon a 
somewhat advanced plinth of two chamfered orders, an indication 
of date that cannot be ignored. As noticed above, ante, p. 273, 
none of the buildings of the first period have plinths, while on 
the other hand these are common in the late western towers 
and churches of the last sub-period (see Fig. 31, ante, p. 85). 
Either then the transepts are late tenth century and the 
tower arch c. 1040, or the transepts are c. 1040 and the 
tower arch Early Norman. 

(2) Some late examples with angle shafts, etc., advanced 
mouldings, or recessed arches : — 

Bosham, Sussex ; Broughton ; Carlton-in-Lindrick, Notts ; 
Clayton, Sussex ; Kirk Hammerton and Kirkdale, Yorks ; 
Norton, Durham ; Sompting, St. Botolph, and Stopham, Sussex ; 
Stow (archivolt of tower arch) ; Wareham ; Wittering. 

On the basis of this chronological survey we may now 
attempt a brief historical sketch, which may serve as a 
summary of previous discussions. 

It has been already seen that though there are monumental 
links between the Early Saxon building period and that of 
Romano-British times (Vol. i, Ch. vii, ad init.), the strictly 
architectural connection is slight. There is no evidence that 
any existing Saxon church was once part of a pagan Roman 
building (ante, p. 125). In each of the cases where this has 


been suggested, form and orientation betoken an ecclesiastical 
origin. Whether or not any part of a Romano-British church 
survive in an existing Saxon one is another question. Sil- 
chester has come down to us, and we have seen some reason 
to believe that the fabric of Augustine's cathedral at Canter- 
bury, enlarged in the tenth century to be destroyed in the 
eleventh, was Romano-British (ante, p. 261). We have every 
reason to credit the assertion of Bede that the original church 
of St. Martin at Canterbury was the work of Roman believers.^ 
There is a parallel to it at the other end of Britain, in Ninian's 
church of St. Martin at Whiterne in Galloway of about the 
year 400 (Vol, i, p. 161), and the two Martin churches are 
links in the chain of evidence that connects the Church in 
Roman Britain rather with Gaul than Italy. The present 
chancel may as we have seen contain some of the earlier work, 
but the fragment is not, like the cathedral and Silchester, 

This fact that Silchester was basilican, while Rochester and 
St. Pancras consisted in single naves (ante, p. 119 f-), is the first 
significant phenomenon in the history of Saxon architecture. 
The width necessitated either a correspondingly spacious arch 
of triumph, or an arcade, and the choice at St, Pancras, 
Rochester, Lyminge, and other places, of the latter is another 
noteworthy appearance. The introduction at the first named 
churchj and perhaps at the cathedral (ante, p. 262), of projecting 
lateral chapels is a third point, and the occurrence at St. 
Pancras, Monkwearmouth, and Corbridge, with perhaps other 
examples, of western porches is a fourth. For none of these 
characteristics or features can we readily find prototypes on the 
Continent. The want of surviving early churches in Gaul may 
be the reason of this, but the literary notices we have of Gallo- 
Roman and early Merovingian churches do not suggest to us 

^H.E. i, 26. The words are worth quoting as they are sometimes mis- 
interpreted, ' erat autem prope ipsam civitatem ad orientem ecclesia in honorem 
sancti Martini antiquitus facta, dum adhuc Romani Brittaniam incolerent.' 


buildings with the characteristics of these Early Saxon examples. 
North Africa and Italy are no more helpful than Gaul, and 
Saxon architecture seems to start from the first on lines of 
considerable originality. The baluster shafts in the north 
(Fig. 82, ante, p. 145) are equally sui generis, or at any rate 
there are no Gallic prototypes. The fact that the earliest work 
at Monkwearmouth is so unlike what Gallic workmen would 
have wrought, is one argument tor placing it rather later than 
the actual time of Benedict Biscop. 

For another phenomenon of the earliest group of churches 
it is easier to find continental parallels. This is the presbyterial 
space screened off before the apse which is so marked an 
internal feature at Brixworth (Fig. 150, ante, p. 249). In a 
rudimentary form we find this in the ' stilting ' of the apse in 
most of the early churches, e.g. St. Pancras and Reculver, 
which have this termination. This is only what occurs with 
some frequency in Early Christian churches outside Italy, 
as in North Africa and Syria, and calls for no special remark. 
The Brixworth arrangement is not, to the writer's knowledge, 
found elsewhere, though the transept of the basilicas of the 
city Rome has been brought into comparison with it. Screens 
formed of arcading are however used to cut off spaces at the 
altar end of interiors in not a few early churches, especially in 
Spain. ^ Among churches that exhibit this feature the nearest 
parallel to Brixworth appears to be the instructive Carolingian 
basilica at Michelstadt (Steinbach) in the Odenwald, built by 
Eginhard about 825 a.d. The interior of this now desecrated 
building preserves clear traces of a cross wall about 12 ft. high, 
which with openings in the centre, formed a screen that cut off 
a space of about 16 ft. from the eastern end ot the nave.^ At 
Brixworth the wall rises to the roof and makes the division more 
marked, but it was probably due to the same reason which 
operated at the later date at Michelstadt. 

1 Dchio u. von Bc/.old, i, 98, enumerate some examples. 

2 Adamy, Die Einhard-BasUlka zu Steinbach im Odenzvalde, Hannover, 1885. 


Like so many of the Saxon churches of the first genera- 
tion, Brixworth was monastic, but was at the same time a 
missionary church to which the surrounding population had 
to be conducted or allured. Michelstadt evidently had the 
same intention.^ The part of the interior cut off at the 
altar end would serve as the monks' church, the general 
congregation occupying the nave. 

The plans of the earliest Saxon churches were as we have 
seen partly apsidal and partly square-ended, the latter pro- 
bably preponderating. As in their plans, so too in their 
technical aspects, Saxon churches exhibit a mingling of the 
two traditions that were noticed in the opening chapters as 
operative in our pre-Conquest building. Saxon masonry, 
commonly of irregular rubble-work, was compacted in walls 
as a rule of remarkable thinness. These thin rubble walls 
are of the same character as the partition walls in Roman 
villas and stations, which in early Saxon times were suffi- 
ciently abundant to supply models in almost every part where 
these were needed. Roman squared stones were used 
wherever available for quoins and special features, and it 
was probably his familiarity with these that gave the Saxon 
builder that penchant for the megalithic which never left 
him. Celtic tradition on the other hand, if not solely 
responsible for the normal nave and chancel plan, makes 
itself apparent in the curious feature of sloping jambs to door 
and window openings. The phenomenon is by no means 
universal, but it appears from time to time through the 
whole course of Saxon architectural history. It is very 
pronounced at the early Escomb (ante, pp. 114, 115) but is 
to be observed also in the aperture of the late chancel window 
at Boarhunt (ante, p. 105) as well as in that at West 
Hampnett, Sussex ; also in openings in the fine church which 
cannot be an early one at Brigstock in Northamptonshire 
(Fig. 43, ante, p. 94). 

^Adamy, loc. cit. p. 6. 


With materials and technique that are for the most part 
Roman, but exhibit some Celtic peculiarities, the Saxon builder 
of the seventh and eighth centuries constructed the monastic 
and village churches the number and distribution of which 
have been already indicated (ante, p. 75 f. and map, Fig. 175). 
A certain originality in planning as well as in technique 
and ornament has been claimed for them from the first, and 
it is probable that the churches of the seventh century were 
fully as ambitious and well executed as those nearer to 
the time of the Danish inroads. There is to be noticed about 
many of the former a certain amplitude, which suits a time of 
ease, and which is marked alike in general proportions 
(Rochester, Brixworth), and in openings (St. Pancras, Cor- 
bridge). The characteristic narrow Saxon doorway, which we 
find at Bradford-on-Avon or at Worth (lateral doors), is rather 
late than early. The internally-splayed high and narrow loops 
as at West Hampnett, Sussex, are also late. 

So soon however as the disastrous and terrifying Danish 
inroads had become the predominant feature of the times the 
art of building must have received a check, for though a 
church ruined by a Viking raid would as a general rule be 
rebuilt, yet as such raids were often repeated there was no 
encouragement for display or elaboration in any new or reno- 
vated fabric. Notwithstanding this, the art of building 
during the second or Danish period was certainly not at 
a standstill, for the development of the special Saxon 
peculiarity of the long-and-short quoin must fall within 
this time. It derives its origin, it will be remembered, from 
some of the earliest work, and it is in normal use in the 
latest period, so that its evolution must fall in the inter- 
mediate epoch. It is not easy however to identify long- 
and-short quoins, so to say, in the making, for this special 
arrangement of pieces may occur accidentally in quoins that 
are not intended to be of this particular character. A more 
minute examination of our Saxon buildings may reveal evidence 


of the gradual formation of their characteristic features, but 
such ' transitional ' forms are at present difficult to identify. 
The quoin shown in Fig, 32, ante, p. 86, may be regarded by 
some as transitional, and the quoins at Sockburn, Durham, 
when compared with Escomb carry the same suggestion/ 

Before the beginning of the intermediate or Danish period 
there was already established that connection between England 
and Germany the importance of which in the domain of 
the arts has already been noticed. This continued through 
the intermediate period, and it is during this that we should 
naturally look for the traces of kinship between Saxon and 
Austrasian architecture. The special features, on which the 
suggestion of this kinship is founded, do not however come 
into vogue in Germany before about the tenth century, and 
we are inclined to regard their introduction into English 
work as due to the marked activity in church building and 
restoration that signalized the reign of Edgar (959-975 a.d.). 
There has been already quoted a significant remark relating to 
the abbey church at Ramsey of c. 970 (ante, p. 242), It is 
said of it that ' compared with the old fashioned method ot 
building which had before prevailed, it was a structure of no 
mean pretension,' and we see proof here of the awakened 
ambitions of the age. The most intelligible theory of the 
architecture of this epoch seems to be that when the new activity 
began the English builders of the time found themselves rather 
at a loss for features which should give an architectural character 
to their fabrics, and were glad to adopt the pilaster strips of 
their neighbours across the North Sea. The substitution of 
the double-splayed lights for the internally-splayed ones 
which were universal in the earliest epoch and appear e.g. 
at Avebury, and at St. Michael, St. Albans, of about 950, must 
have been due to the force of the new influence, which we 
regard here as making itself felt all at once at this epoch of 

^ The Reliquary^ April, 1894. 


revival, rather than as slowly filtering in through a long period 
of years. 

Whether or not the western tower itself, with its equip- 
ment of double belfry openings, made its appearance in the 
same comparatively sudden fashion is a question on which 
a word must be said. There are here certain con- 
siderations on which it is worth while for a moment to 
dwell. The western tower in Saxon England is in some 
cases connected with the western porch, and this last is an 
early feature. The western tower is moreover associated 
with subsidiary structures which have left their traces at 
Brixworth, Corbridge, Netheravon, 
and other sites ; while at Brixworth 
and Deerhurst it has chambers within 
it evidently destined for some pur- 
pose of honour. That is to say the 
Saxon western tower has about it 
features which give it a special position 
and interest, and oblige us to consider 
it as an independent creation. Now F'^- i^g.-Plan of western 
, . . , J • J tower at Deerhurst. 

there is one exceptional tower, devoid 

of any of the later features, that may well be earlier than 
Edgar's time, and that is attached to a church which also 
presents no details of acknowledged lateness. The reference 
is to Deerhurst, the scheme of the eastern part of which 
was given in Fig. 140, ante, p. 231. 

The western tower at Deerhurst, of which Fig. 169 shows 
the plan, is unique in that it is divided into two by a parti- 
tion wall that runs up to about a third of its present height 
of 71 ft. Externally it is perfectly plain without string course 
or set off. The belfry openings with practically the whole 
of the upper part of the tower are of later date. The 
internal arrangements were noticed on page 171 f., 
and it will be remembered that the chamber containing 
the aumbry-like recesses in the middle third ot the 



tower opens towards the nave by a double aperture. 
The form and details of this are of much significance and 
are shown in Fig. 170. It must be noted that the pier which 
divides the aperture runs through the whole thickness of 
the wall and is not a mid-wall shaft. Furthermore it is 
enriched with flutings, that are alternately in their upper or 
lower halves filled in witli convex members in the fashion called 
* cabling,' while all the details are cut with classic decision. 

Fig. 170. — Double aperture in eastern face of Deerhurst tower, western side. 

No such treatment of double openings and no enrichment 
of jthe same pattern occurs elsewhere in our Saxon work.^ 
We may indeed single out the feature in question as the 
only one in any Saxon building that can be distinctly fixed 
as transitional between the first and the third of our periods. 
We find certain forms in evidence in the earliest groups 
and others in vogue in the latest period, but we have just 

1 The reeded pilasters at Bradford-on-Avon are much coarser attempts. 


seen how hard it is to find any link of connection between 
the two sets. This case of the subdivision of an opening 
by means that are not yet those of the normal mid-wall 
shaft and through-stone, is accordingly of especial significance 
for the chronology of Saxon work. A case not wholly 
unlike at Earls Barton has been noticed ante, p. 189 f. In 
another connection the character of the detail has equal 
significance. Carolingian art supplies us here with some- 
what close parallels. On the western face of the northern 
half-round flanking turret of the western forebuilding at 
Aachen occurs a small opening, divided by a square fluted 
pilaster, that can be distinguished in Fig. 20, ante, p. 51. 
Similar pilasters, that are modern restorations, divide openings 
in Carolingian walling that abuts on the north western 
corner of the minster, and fluting of the same classical 
pattern occurs elsewhere in Carolingian work. Some capitals 
from the palace of that period at Ingelheim, in the Museum 
at Mainz, show it, and so do certain imposts in the church 
at Hochst on the Main, placed by Essenwein in the ninth 
century.^ The ' Vorhalle ' at Lorsch in its upper stage 
(Fig. 22, ante, p. 59) also furnishes an instructive parallel. 

It is not necessary to suppose Deerhurst Carolingian, but 
the presence of these early details, and the absence both 
from tower and church of the recognizable marks of the 
later period, seem to claim for it a place apart. Let us 
therefore test this hypothesis of a date in the intermediate 
or Danish period by examining the church as a whole and 
by a comparison with other buildings that are possibly of 
the same age. 

Deerhurst was a monastery.^ The exact date of its founda- 
tion is not known but it was in existence during the Danish 

'^Handbuch der Architecture die Baustile, in, p. 139. 

2 For the history and description of Deerhurst see the excellent monograph 
by the Rev. G. Butterworth, Deerhurst, a Parish of the Vale of Gloucester, 
Tewkesbury, William North, 1890. 


period. The arrangements of the eastern end of the church 
would concern the inmates of the monastery, the buildings of 
which, at any rate in later mediaeval times, adjoined the south- 
eastern part of the edifice. Over the transeptal chapels seen on 
the plan, Fig. 140, ante, p. 231, there were upper chambers that 
opened towards the choir by wide arches. These apertures 
with the various doorways on the ground floor, which have 
square, triangular or round heads, are treated with great 
simplicity. There are no flanking pilasters but the imposts in 
some cases have a hollow chamfer. The most advanced feature, 
one that occurs also in the tower, is the square sectioned hood 
mould which we find over the wide arch leading towards the 
apse. This, like hood moulds on the tower, springs from pro- 
jecting corbels in the shape of animals' heads. Such carved 
grotesques suggest a Scandinavian influence that might easily 
have been exercised in this period, and this suggestion is borne 
out by the appearance at a later date of similar sculptured 
heads on the Norman church of Kilpeck in Herefordshire, in 
proximity to ornamental pilasters carved with intertwined 
serpents of a pronounced Scandinavian type. Save for the 
hood moulds which introduce an element of doubt, there 
seems nothing in the forms or details at Deerhurst incompatible 
with an attribution to the early part of the tenth century, 
while the narrow doors to the transeptal chapels on the ground 
story are also early. The projecting half-rounds on the jambs 
of the arch before the apse are not ordinary soffit shafts, and the 
arch above them is not recessed or furnished with a soffit roll. 

The arrangements of the western end suggest a comparison 
with Brixworth. The apartment there with the triple window 
on the first floor of the tower is a parallel to the apartment 
with the double opening at Deerhurst. Just as at Brixworth 
access is provided to this by the turret stair, so at the latter 
place the doorway in the western face of the tower on the level 
of the aforesaid apartment (ante, p. 1 74) may have been gained 
from outside by some convenient arrangement for ascent. 


Who made use of these apartments ? There is a treatise by 
the Carolingian statesman Eginhard, in which he tells of 
certain miracles wrouo-ht in the basilica he had erected about 
830 at Seligenstadt, then Miihlheim, on the Main, In the 
upper story of the western choir he possessed what he calls 
a ' coenaculum ' or upper chamber in which was an altar and 
which he used for his own accommodation during the services.'- 
When we reflect on the position assumed towards the churches 
of the later Saxon period by the great men of their localities 
(see Vol. I, p. 324 f.) we may picture to ourselves the local 
landowner at Deerhurst or Brixworth following the example 
of the Carolingian statesman and establishing himself in a 
coenaculum at the western end of the local oratory, after the 
fashion of an English lord or squire of much more recent days. 
Brixworth was at first monastic and Deerhurst was monastic to 
the end, but a small monastery often depended on the local 
lord, and in any case the western end of its church was 
generally for the use of the lay population. Brixworth may 
have been restored by the ealderman of the place even prior to 
the time of Edgar. The Saxon work of the restoration shows 
none ot the often-noticed later features, and the balusters in 
the opening have behind them an old Saxon tradition (Fig. 
1 14, ante, p. 198). 

This suggestion of the use of these upper chambers by some 
local magnate cannot be dissociated from the question of the 
purposes of the various western chambers and adjuncts con- 
nected on the ground story with towers, of which the 
monuments have given us evidence. In many cases no doubt 
these were set apart for the administration of baptism, and 
that they were also used for burial we can gather from what 
Bede tells us of the porticus ingressus at Monkwearmouth-as well 
as from the story about the sepulchre of Swithun at Winchester.^ 

1 Mignc, Pat. Curs. Compl. civ, pp. 559, 565, 593. 

2 Historia Abbatum, c. 20. 

^'Winchester' Volume of the Archaeological Institute, p. 6. 


We can however do little here but conjecture. The existence 
of these various adjuncts and coenacula gives an importance to 
western ends of Saxon churches which bears out the theory of 
their connection with the Austrasian ' Westwerke ' rather than 
with simpler frontal towers of defence such as we have seen 
existing in parts of France. Taken by itself Deerhurst tower 
might suggest a structure of this latter kind, and the site of the 
monastery near the Severn, v/hose waters were so often 
ploughed by Viking keels, would make this a plausible 
theory of its origin and character. Viewed as a whole, how- 
ever, Saxon towers look much more like the open, freely-used, 
German forebuildings than like closed towers of defence. 
Deerhurst had a western doorway and about half of the other 
Saxon towers possess this feature, while all, save Leathly, ante, 
p. 1 66, open to the church through an ample tower arch. 

Deerhurst and perhaps Brixworth may accordingly be 
regarded as introducing into our architecture the western 
tower, and their towers may even be claimed as some of the 
earliest of existing specimens in Europe generally, preceding in 
point of time the numerous western towers of the Rhineland 
and Westphalia. In relation to the cruciform plan also Deer- 
hurst may mark an epoch, for the transeptal chapels, as we find 
them here with their two stories, are bold features that would 
tell in the external view, though internally they only open 
towards the central space through narrow doorways. 

The third period opening with the reign of Edgar claims by 
far the greatest number of the extant monuments, and this fact 
gives it a proportionate importance. The same nave and 
chancel plan which we find at the earliest epoch remains in 
common use to the end of the whole Saxon period, and is then 
taken over as an inheritance by the Norman and the later 
mediaeval builders. This normal scheme is however very 
commonly extended by the addition of a western tower, which 
is established in this period as a substantial feature of English 


church architecture ; while axial towers in various positions, 
towers forming the body of a church, twin-towered facades, 
and central towers connected with the cruciform plan or at 
times combined with western towers, supply interesting material 
to the student of architectural evolution. Side by side with 
these square-ended single-aisled churches, the period gives 
us at Wing one developed basilican plan with a polygonal 
apse and this may probably be regarded as for this country 
the expiring effort of Early Christian art. 

There can be no attempt here at any detailed analysis, from 
the historical side, of the copious material offered by this 
ultimate period. All that can be done is to glance at the 
principal groups and types from the standpoint of the scheme of 
chronology offered a few pages back. 

The fact that as a group the ' Lincolnshire ' towers must 
almost necessarily be very late has an obvious bearing on the 
suggestion that they had a defensive character. It has been 
noticed already (ante, p. 166) that there is nothing about the 
towers themselves to indicate such a character, but to many 
people their occurrence in such numbers in regions that were 
specially exposed to Danish visitations will seem hardly 
fortuitous. Ecclesiastical towers were certainly used in other 
lands for purposes of defence and refuge. The Irish round 
towers are the most conspicuous instance, but many of the 
single frontal towers in France and even some in Germany 
were of this kind.^ In our own country near the Scottish 
march there are church towers which served a quasi-military 
purpose in Border warfare. In the case of the Irish round 
towers, it is possible that those which actually remain are not. 
the earliest which were built, but that ruder structures of the 
same kind preceded them. So too it is conceivable that the 
* Lincolnshire ' towers of the eleventh century represent a type 
evolved somewhat earlier when Danish hostility was a still 
present danger. It has just been admitted that defence or 

^Dehio, Kirchliche Baukunst, i, 586. Enlart, Ma?iuel, p. 249. 


refuge may have had something to do with the planning of 
Deerhurst tower, and it has seemed best to indicate this theory 
in the case of the ' Lincolnshire ' group, though in the view of 
the writer the theory in question has no grounds, beyond a 
certain general likelihood, for its support. 

The tower in its other aspects as axial and central has been 
ascribed to this period, and the full development of the cruciform 
plan waited too for the same epoch. At Deerhurst, where the 
transepts first seem to assert themselves on the exterior, 
there is no evidence of a central tower, and Barton-on- 
Humber may claim to be the earliest tower of the kind. The 
continental plan most like that of Barton, at Werden a.d. 
Ruhr, seems to have consisted in a square central tower 
about 30 feet on a side, in interior m.easurement, with aisles 
to west, north and south, the east side being joined on to an 
earlier church.-^ The arrangement is that of a ' central ' church 
and may have been suggested by the Early Christian Greek- 
cross plan with central pavilion-tower. Similarly Barton-on- 
Humber may really rest on a like tradition. There is no 
reason to suppose it copied directly from Werden, or any 
similar continental example, though it is of interest to know 
that schemes of the sort were in use in the region with which 
the designer of Barton is most likely to have been in touch. 
The Greek cross plan existed already at Hexham, and Barton 
may be in indirect fashion affiliated thereto. At Barton the 
tower, it will be noted, has full independence, and it is possible 
that this scheme came ultimately to influence the designers 
of the completely cruciform churches, such as Norton and 
Stow, in which the tower has the same distinct individuality 
apart from the nave and the other members that abut against it. 

Wilfrid's church would then have established the principle 

of the central tower, and Barton have expressed the same idea 

in the language of nascent Romanesque, while Stow and 

Norton perfected the scheme in connection with the Latin 

^ Effmann, Die Karol.-Oiion. Bauten zu Werden^ p. 168. 









cross which by the middle of the eleventh century was becom- 
ing obligatory. Meanwhile quite a distinct line of develop- 
ment was being followed in those examples where we see the 
tower growing out ot the nave walls, and the transepts 
gradually forming themselves out of the smaller side chapels 
that were an inheritance from the earliest times. The dates, as 
already indicated on the evidence of details (ante, p. 281 f.), 
would agree with this. Britford, Deerhurst, Repton, are early 
in, or even prior to, the third period, and show the transept 
still only a side chapel without any evidence of a tower. 
Breamore, also Vv^ithout specially late indications, shows the 
growth of both transept and tower, and Worth and Dover their 
fuller development (Fig. 171).^ Barton has already established 
the tower as partly central, and Peterborough, and perhaps 
Stow, the transepts though only on a crux commissa scheme. 
For the final outcome of the process of evolution we wait till 
the latest division of the period, when Stow in its later 
form and Norton present us with harmoniously combined 
schemes in which church architecture attains to fairly complete 

^ Fig. 171 is based on a plan in the Irvine drawings. The church is 
sometimes figured without any indication of transeptal arches, but with the 
transepts quite open as at Peterborough. The existence of a central tower 
shows however that there must have been walls here with openings into the 
transepts, otherwise the tower could not have been carried. The width of 
the Saxon openings is not known. The present transeptal arches are of the 
twelfth century. 



The preceding four chapters have been occupied with a 
description and analysis of the existing Saxon monuments, 
followed by some necessarily tentative notes on the chronology 
and history of the style. It remains now to attempt in a few 
words a critical estimate of pre-Conquest buildings from the 
point of view of the general history of the architectural art. 

Hitherto the study has been chiefly of extant monuments, 
and only occasional reference has been made to examples known 
only from literary sources. To obtain a critical estimate, 
account must be taken of the buildings that have perished, as 
well as of those which in part survive. 

As was noticed on a previous page (ante, p. 72), the great 
majority of existing remains are those of churches of the 
' village ' or ' parish ' type, though fragments exist of at least 
two bishop's churches, Rochester and Sherborne, and of one 
abbey church of an important establishment, at Peterborough. 
Besides these there are several monastic churches of lesser rank, 
such as Oxford (St. Frideswide, later the cathedral), Lyminge 
and Reculver, Monkwearmouth and Jarrow, and Deerhurst ; 
as well as collegiate churches of secular priests, for this was the 
status of Stow after 1040, and may have been that of other 
examples. Most of these monastic and collegiate fanes de- 
scended afterwards to the rank of parish churches, in which 


character they have come down to us, and it may be said 
generally that, so far as size and architectural pretension are 
concerned, the larger village churches of Saxon times were on 
the average quite equal to those of monastic or collegiate rank. 

On pp. 104, 105, 107, no, 184, 185, 210, are plans of 
churches of the ' parish ' type, most of which would be pro- 
nounced below the later average of size, but on the other hand 
Dunham Magna, p. 225, Britford, p. 228, and St. Martin, 
Canterbury, p. 120, are normal buildings of their order, 
measuring 43 ft., 44 ft., and 38 ft. 6 in. respectively, in length 
of nave, while Breamore, p. 233 (75 ft.), Worth, p. 237 
(59 ft.), Wing, p. 268, Bosham, postea, p. 328 {^6 ft.), and 
Repton are decidedly large. The last-named, which was after 
the Danish troubles only parochial, corresponds in size as well 
as in general scheme of plan with the monastic Deerhurst. 
(See ante, p. 231.) 

To judge by surviving examples the Saxon village church of 
stone, though architecturally plain, was a building not far below 
the average size and pretension of a village church of the later 
mediaeval period. The dimensions of the naves of about a 
score of the best preserved Saxon examples both large and small 
figured on a former page (ante, p. 84) will be found to give an 
average not very different from that of the naves of twenty Nor- 
man or Early English examples, chosen as fair specimens of their 
classes. The nave, which sufficed for the congregation of the 
eleventh century, has, in most cases but not in all, received the 
addition of a north and it may be also a south side-aisle, while 
the chancel has been enlarged eastward to suit a more elaborate 
ritual. These additions merely however correspond to the 
gradual increase in population, which went on till the visitation 
of the Black Death in the fourteenth century. They were on 
the same grade of architectural pretension as the rest of the 
fabric, the general character of which remained unaltered. 

The Saxon village church was therefore fairly up to the 
general mediaeval standard for structures of the kind. Can we 


say the same of the more imposing edifices ? Did the Saxon 
cathedral and abbey churches, taking these also on the 
average, reach the standard of the later Norman and Gothic 
fanes ? The right answer would probably be that while 
in these important structures the Saxon builder rose fairly 
to the height of his task, yet the standard of size and splendour 
mounted, in their case, so rapidly after the Norman Conquest, 
that the original Saxon structures had perforce to yield to 
larger and more ornate edifices. From the tenth century to 
the middle of the fourteenth, the economic condition of the 
rural villages and country towns did not greatly alter, whereas 
the resources of the greater abbeys and bishops' sees were 
enormously increased. From the fiscal point of view, as Mr. 
Prior has pointed out,^ striking results followed from the 
exploitation of the shrines of saints and martyrs, to which 
pilgrims, with offerings in their hands, flocked in crowds from 
far and near. The immense store of riches thus accumulated 
at the shrines of Cuthbert or Becket or Edward II enabled the 
abbots and bishops of Durham, Canterbury or Gloucester to 
build and to decorate as lavishly as they might desire, while 
such a use of their funds was forced upon them by the general 
tendency of the age. 

The age which opens with the settlement of the Normans in 
England was in fact one of immense architectural activity. It was 
the ambition of each generation of builders to surpass all that 
had been done before, and the Normans of the second period 
treated the earlier efforts of their countrymen with the same 
scant courtesy that was shown at the Conquest to the Saxon 
edifices. It is a striking fact that though Archbishop Lanfranc 
had erected from the foundations at Canterbury a new Metro- 
politan church that must, at least, have equalled the great 
abbeys at Caen, yet, within a generation, the priors Ernulph 
and Conrad pulled down his choir and re-erected it on a scale 
of transcendcntly greater magnificence. The Norman abbey 
'^History of Gothic Art in England^ Lond. 1900, p. 166. 


church at Peterborough is more than double the size of the 
Saxon structure burnt down in 1116, the plan of which has 
partially been recovered (Fig. 172 postea, p, 315), but the 
' glorious choir ' of Conrad, at Canterbury, covered about four 
times the space of that occupied by the eastern part of the 
cathedral of Lanfranc. 

It is no reflection, therefore, on the intrinsic character of 
the Saxon cathedral and abbey churches that they were all 
replaced after the Conquest by grander structures. It may 
be interesting to ask. the question — When, and in what 
circumstances, did this substitution take place ? 

It is worthy of note that, of the important Saxon churches, 
the one which lasted longest into the mediaeval period was 
among the earliest of all. At Hexham, it appears that Wilfrid's 
edifice of about 675 was still serving as the nave of the 
extended abbey church up to the devastation wrought by the 
Scots in 1296.^ None of the others survived so long, but 
we need not assume that the Normans condemned them at 
once as intrinsicallv v/orthless. In the case of five of the 
most important cathedrals and abbeys, Canterbury, York, 
London, Gloucester, and Peterborough, and in that of Here- 
ford also, the Saxon buildings were ruined by fire within a 
few years of the Conquest, and had necessarily to be rebuilt ; 
while Rochester, Wells, and St. Frideswide's, Oxford, were in a 
greatly dilapidated condition. At St. Albans, at the end of 
the tenth century, a Saxon abbot set on foot, though he 
did not carry out, a grand scheme of re-building, which 
points to the fact that the actual structure was recognized 
as small or faulty. At the important sites of Winchester, 
Ely, Durham, Worcester, Exeter, the re-building seems to 
have been deliberate. Worcester was re-constructed by its 
Saxon abbot Wulfstan, who pulled down the structures of 

1 The Chronicle of Lanercost (Bannatyne Club ed., pp. 1 74-5), ad ann. 
1296, tells us that in that year ' ipsa vero basilica Romano opere insignita ad 
honorem . . . Sancti Andreae . . . beati Wilfridi ministerio exstitit dedicata.* 


his predecessor, Oswald, because they were not large enough, 
though he shed tears at an act which seemed to him to 
savour of sacrilege.^ At Exeter, the first Norman bishop 
was thought to show a certain want of spirit and ambition, 
in that he was content with the ancient buildings of the 
abbey ascribed originally to iEthelstan and restored after the 
Danish wars by Cnut. An ancient seal, attached to some 
charters in the possession of the Exeter Dean and Chapter, 
is believed to give a representation of the facade of this 
edifice (see Fig. 147 ante, p. 243). Such as it was however, 
the second Norman bishop, Warelwast, was not satisfied v.-ith 
it, and commenced in 1 1 1 2 a new pile of which the well- 
known transept towers are surviving portions. At Winchester 
and at Ely the Normans found on the sites buildings dating 
in all essentials from the great era of church restoration in 
the latter part of the tenth century. Such buildings might 
have stood for hundreds of years, but they were completely 
removed and new edifices constructed from the foundations 
by the first Norman bishop and abbot. 

The facts here noted and the exclamation of Wulfstan of 
Worcester, ' Wretches that we are, we destroy the work of 
our saintly forbears because we think in our pride that 
we can do better,' testify to the new ambitions which 
inspired the greater architectural undertakings after the Con- 
quest, but still allow us to credit the Saxon fanes with some 
nobility of proportions and workmanship. They had been 
in their day fully sufficient for their purpose, Wilfrid's 
church at Hexham, even in the twelfth century, excited the 
admiration of the contemporaries of William of Malmesbury. 
The writer just mentioned understood architecture, and we 
owe to him some instructive critical remarks. It is he who 
notes the novelty of the style of the Confessor's church at 
Westminster, the first example in England of Norman 

^ Nos, inquit, miseri Sanctorum opera destruimus ut nobis laudcm com- 
paremus. W. of Malmes. Gesta Pontijicum, Rolls Series, No. 52, p. 283. 


building,' and signalizes the fine jointed Norman masonry- 
employed by Bishop Roger of Salisbury.^ It means some- 
thins: therefore when he tells us that those who had visited 
Italy seemed at Hexham to see the glories of Rome revived 
before their eyes, and himself remarks of the Saxon church at 
Malmesbury, either Aldhelm's or a reconstruction of the tenth 
century, that the whole fabric remained untouched and 
conspicuous to his own time, surpassing in beauty and in size 
every ancient building that was to be seen in England.^ At 
the beginning of the twelfth century Edmer, who had been 
to Rome with Anselm, writes of the Saxon cathedral of 
Canterbury, begun by Augustine and enlarged by Archbishop 
Odo about 950, as having been built partly in imitation of 
the great Roman church of the prince of the Apostles. The 
Saxon cathedral at Durham, erected at the end of the tenth 
century, was, according to Simeon of Durham who had seen it, 
a stone church of fair appearance and magnitude (honesto 2iec 
parvo opere) but it was replaced a hundred years after by 
a Norman edifice larger and of grander show (nobiliori satis et 
majori opere). ^ 

Such testimony from intelligent writers familiar with the 
achievements of Norman architecture justifies us in assuming 
that the Saxon cathedrals and abbey churches, more especially 
v/hen a building was cathedral and abbey church in one, would 
fairly have held their own with continental monuments of the 
same period. These cathedral-abbeys and the other Bene- 
dictine houses of the first rank belonged to wealthy and 
pushing communities, and such monastic churches as those of 
Winchester, Peterborough, St. Augustine and Christ Church, 
Canterbury, St. Alban, Glastonbury, or Gloucester, would have 
exhibited the best architecture of which the time and country 
were capable. We can only test this monumentally in the case 

1 Gest. Reg. ad ann. 1066. ^ ibid, ad ann. 1 1 19. 

^Gest. Pont. loc. cit., p. 361. 

^Hist. Dunelm. Eccl. Rolls Series, No. 75/1, p. 81. 



Fig. 172. — Foundations of castcMi p:iit of Saxon abbey church at 



of Peterborough, where the plan of the eastern part of the 
Saxon abbey church has been preserved (Fig. 172). From 
what exists we can judge of nothing more than the shape 
and size of the edifice. The former suggests the Early 
Christian T form, and seems to place it in the seventh century. 
The latter, which should be compared with the size of the 
other plans, all on the same scale, given in this volume, seems 
decidedly imposing. The width across the transepts was 92 
feet. The length of the nave is not known. The large scale 
of this early monastic church enables us to understand the 
remarkable size of the daughter establishment at Brixworth 
(Fig. 151 ante, p. 248). The imposing dimensions of Brix- 
worth make it all the more likely that these relics of the 
mother church are of the same early period, while the transept 
at Peterborough may account for the presbyterial space 
arranged for at Brixworth before the apse. 

At the time therefore of the Norman Conquest, it may be 
assumed that the chief ecclesiastical sites were supplied with 
buildings corresponding to the demands of the age. The fact 
that in every case these Saxon edifices have yielded place to 
later structures may be taken to show that, though sufficient 
in their time, they were not up to the later mediaeval standard 
in size, solidity, or architectural character. 

Passing now from the question of the scale of the monu- 
ments, to that of their planning, we may ask whether Saxon 
buildings simply reproduced the standard patterns found else- 
where or showed evidence of boldness and novelty in archi- 
tectural design. Even about our churches of the smaller type 
we have found such evidence. The disposition of the porch 
in its different situations ; the growth of the lateral porch into 
the transeptal chapel, and of the transeptal chapel, through 
the widening of its door, into the transept, is a piece of 
architectural evolution which shows a certain originality and 
independence in the Saxon builders. The position of the 


characteristic north and south doorways, the erection of a tower 
over an earlier porch, the western axial tower in its relation 
to western divisions in naves, the growth of the central tower 
and its ultimate independence of the parts that abut against it, 
are points of interest that give life and individuality to the 
story of our early mediaeval building. It is true that these and 
similar innovations, the list of which it would be easy to 
extend, might be put down as mere insular peculiarities with 
no architectural significance of a general kind. Apart from 
these however, there are facts attested by literary records which 
seem to exhibit certain Saxon monuments as distinct landmarks 
in the architectural history of the west. 

From this point of view the accounts we possess of the 
church erected by Wilfrid at Hexham between 672 and 678 
A.D.-^ are among the most important documents that exist about 
the architecture of the seventh century, an era when the 
transition from Early Christian to Romanesque forms was 
already in progress. The fullest account we owe to Prior 
Richard of Hexham in the twelfth century,^ who gives us the 
impression of so elaborate a structure that, if his notice stood 
alone, we should conclude it referred to some later rebuilding 
rather than to the actual work of Wilfrid. Wilfrid's own 
choirmaster Eddius however describes the building in almost 
the same terms, and we obtain in this way contemporary 
evidence for features we should certainly not expect to find 
together in this remote region and at so early a date. Eddius 
tells us that the building in its lower parts contained certain 
chambers in the earth wrought of well-polished stones, while 
above ground it was of many parts, supported by numerous 
columns and side aisles or chapels (columnis variis et porticibus 
multis sufFultam) while the walls were of notable length and 
height. It had lines of passages with many windings that led 

^ The chronological indications for the dates of Wilfrid's churches at Ripon 
and Hexham are given in Plummer's Bedc, 11, 318. 
^Twisden, Decern Scriptores, Lend,, 1652, col. 290. 


sometimes up and sometimes down and communicated by 
winding stairs.-^ Prior Richard adds the important notes that 
the walls were in three stories (tribus tabulatis distinctos), that 
there was an ' arch of triumph ' or chancel arch (' arcum 
sanctuarii ' which does not necessarily imply an apse), and that 
this arch, together with the capitals of the columns, was 
enriched with carvings in relief (variis celaturarum figuris ex 
lapide prominentibus). Furthermore that there were many 
oratories in the aisles or chapels (in ipsis porticibus) with 
sundry altars of the Virgin, St. Michael, St. John, and the 
holy apostles, martyrs, confessors and virgins, some of which 
to the writer's own day seemed still to rise like towers or 
outworks above the rest of the fabric. All these indications 
bear out the appellation ' multiplex domus ' given to it by 
Eddius, and incline us to credit him when he concludes by 
saying that no such edifice had up to his time been heard 
of on this side the Alps.^ 

'^Historians of the Chunk of 7 ork. Rolls Series, No. 71/1, p. 33. 

2 The two passages referred to above are of such importance that the reader 
may be glad to possess them in the original. That from Eddius runs as 
follows : — Nam in Aegustaldesae . . . domum Domino in honorem Sancti 
Andreae Apostoli fabrefactam fundavit ; cujus profunditatem in terra cum 
domibus mire politis lapidibus fundatam, et super terram multiplicem domum 
columnis variis et porticibus multis suffultam, mirabilique longitudine et 
altitudine murorum ornatam, et variis liniarum anfractibus viarum, aliquando 
sursum aliquando deorsum per cochleas circumductam, non est meae parvitatis 
hoc sermone explicare ; quod sanctus pontifex noster, a Spiritu Dei doctus, 
opera facere excogitavit ; neque enim ullam domum aliam citra Alpes montes 
talem aedificatam audivimus . . . ornamenta hujus multiplicis domus de auro 
et argento lapidibusque pretiosis. . . . 

. The passage from Prior Richard's account of Hexham church is in part an 
amplification of what Eddius has written, but contains independent architec- 
tural statements of value. It belongs to about the middle of the twelfth 
century when, as we are told (ante, p. 312), Wilfrid's church was still 

Descriptio Hagustaldensis Ecclesiae. Igitur profunditatem ipsius ecclesiae 
criptis et oratoriis subterraneis, et viarum anfractibus, inferius cum magna 


At Hexham moreover as we have seen, at a somewhat later 
date, Wilfrid erected, or at any rate began, another church, 
dedicated to St. Mary, that had apparently the plan of a 
Greek cross and consisted in a central part which rose in 
a rounded shape like a tower, and four projecting portions 
on the four sides. ^ The importance of this building, and of 
its central tower-like feature, perhaps an octagon, in connection 
with cruciform plans in general, was noticed on p. 286. 

Another historically important structure of somewhat later 
date was the church at York, rebuilt about the middle of the 
eighth century and described in verses attributed to the 
famous Alcuin. Alcuin was well acquainted with the best 
that Carolingian architecture had achieved in the Prankish 
domains, and his praise of the church is therefore of value. 

industria fundavit. Parietes autem quadratls, et variis, et bene politis columnis 
suffultos, et tribus tabulatis distinctos immensae longitudinis et altitudinis erexit. 
Ipsos etiam et capitella columnarum quibus sustentantur, et arcum sanctuarii 
historlis, et imaginibus, et variis celaturarum figuris ex lapide prominentibus, et 
picturarum et colorum grata varietate mirabilique decore decoravit. Ipsum 
quoque corpus ecclesiae appenticiis et porticibus undique circumcinxit, quae 
miro et inexplicablli artificio per parietes, et cocleas inferius et superius 
distinxit. In ipsis vero cocleis et super ipsas ascensoria ex lapide et deambula- 
toria et varies viarum anfractus modo sursum modo deorsum artificiosissime ita 
machinari fecit, ut innumera hominum multitudo ibi existere, et ipsum corpus 
ecclesiae circumdare possit cum a nemine tamen infra in ea existentium videri 

Oratoria quoque cum plurima superius et inferius secretissima et pulcherrima 
in ipsis porticibus cum maxima diligentia et cautela constituit, in quibus altaria 
in honore beatae dei genetricis semperque virginis Mariae et Sancti Michaelis 
archangeli sanctique Johannis Baptistae et sanctorum Apostolorum, Martyrura, 
Confessorum atque virginum cum eorum apparatibus honestissime praeparari 
fecit. Unde etiam usque hodie quaedam illorum ut turres et propugnacula 
supereminent. . . . 

Atrium quoque templi magnae spissitudinis et fortitudinis muro circum- 

^ (Ecclesia) mirandi operis, et ipsa scilicet in modum turris erecta et fere 
rotunda, a quatuor partibus totidem porticus habens, in honorem Sanctae 
Mariae semper virginis dedicata. Twisden, loc. cit. col. 291. 


The lofty roof of it, he tells us, was upborne by massive 
columns, and aisles and chapels flanked the main edifice and 
contained in all no fewer than thirty altars.^ The notice of 
the building of Ramsey abbey church, c, 970, has already been 
placed before the reader (ante, p. 241 f), while from the same 
important building epoch of the end of the tenth century 
we have an elaborate but confused account of the new fabric 
of the Old Minster at Winchester.^ This seems to have 
been an extensive work, but in the description no special 
feature of historical moment seems to present itself. Hexham 
and York, on the other hand, are places to be noted on 
any map that illustrates the development of ecclesiastical 
architecture in the West. The work done there about 675 
and 780 appears to have had distinct originality and boldness, 
and Hexham, at any rate is a landmark of architectural 

In order to estimate the importance in western architecture 
of Wilfrid's buildings, it will be convenient to glance for a 
moment at the general course of evolution by which the 
mediaeval styles were formed. In the Early Christian archi- 
tecture of the West the one standard form for churches 
used for congregational assembly was the basilica, though 
it was by no means the only form known to Christian 
builders. Now this Early Christian basilica has been made to 
do more than its fair share of work in the development of 
ecclesiastical architecture. It has been assumed that the basilica 
and a few centuries of time were all that was needed for 

^ Haec nimis alta domus solidis sufFulta columnis, 
Suppositae quae stant curvatis arcubus, intus 
Emicat egregiis laquearibus atque fenestris, 
Pulchraque porticibus fulget circumdata multis, 
Plurima diversis retinens solaria tectis, 
Quae triginta tenet variis ornatibus aras. 

Historians of the Church of Tork^ loc. cit, p. 394. 

2 Commented on by Professor Willis in the 'Winchester' volume of 
the Archaeological Institute, Lond. 1845. 


the production of the Romanesque church of the later middle 
ages, but the truth is, that the basilica, when taken by itself,^ 
was strangely lacking in the necessary principle of growth. 
The Romanesque church derived from the basilica the main 
scheme of its rectangular plan, its division into nave and 
aisles and clearstory lighting, and its apsidal termination, 
but for its other chief characteristics, such as stone vaults, 
the use of pillars instead of columns in arcades, a choir 
as the extension of a nave, the central pavilion or tower, 
galleries over side aisles, facades composed with a tower 
or towers, and the like, we have to look to other buildings 
than the basilicas. It is not too much to say that the supposed 
progressive modification of basilican forms, by which Romanesque 
architecture is sometimes explained, is really a figment of the 
imagination. The Romanesque style depended for its forma- 
tion, not on the modification of the basilica from within, but on 
the grafting on the simple basilican scheme of more elaborate 
architectural features that originated elsewhere. The source 
of these is to be sought in the round or polygonal and the 
cruciform churches, that were erected from the earliest times, 
not always for worship, but more often for memorial or sepul- 
chral purposes, or simply as baptistries. These exceptional 
buildings show from the first far more architectural character 
than the basilicas. Their plans are more complicated, but at 
the same time more compact, their technique more advanced, 
their construction more daring and masterly. 

Buildings like San Lorenzo at Milan, San Vitale at Ravenna, 
and Charles the Great's octagonal church at Aachen, contain 
all the constructive and artistic elements that in different 
combinations make up the Romanesque church. They exhibit, 
first, a central space with a side (or rather concentric) aisle, 
vaulted itself and carrying a vaulted gallery opening into 
the central space, which is also covered with a dome in 
masonry. Access to the gallery is gained by stairs in stair- 

^ Ante, p. 14. 



turrets, and these, at San Vitale and at Aachen, are grouped 
on each side of an entrance-porch, thus prefiguring the most 
characteristic feature of the Romanesque elevation. At both 
these places we find even that special mark of advancing Roman- 
esque, the rectangular choir or chancel preceding the apse. 

Here there are most of the characteristic features of 
Romanesque architecture, and all that the mediaeval builders 
had to do was to combine these with the rectangular plan, 
which was the contribution of the basilica. When and where 
the necessary steps were taken it is impossible, in the present 
state of knowledge, to say. It is clear that the basilica in 
itself contained no principle of growth, or we should be able 
to trace the beginnings of Romanesque at Rome and at 
Ravenna. At Ravenna however the basilica maintains 
throughout its simplest fundamental form of nave and side 
aisles, with semicircular and unstilted apse opening directly 
into the former, and the latter unprovided with galleries. At 
Rome certain innovations on the bald basilican form made 
their appearance, but in such a way as to show that they 
were rather accidents than stages in a progress. 

It is curious indeed to note how unprolific in new archi- 
tectural forms were the builders of Rome herself during the 
early mediaeval period. Rome, the mother of ecclesiastical 
statescraft, was in matters of art the most unproductive of 
all the centres of the West. As her architects had constructed 
in the fourth century so they continued to build in the twelfth 
and thirteenth, and after one solitary attempt at Gothic at 
Sta. Maria sopra Minerva, the Early Christian style passed 
at once into that of the Renaissance. This is a fact 
to be remembered when we find mediaeval writers speaking 
of Rome as a place of architectural inspiration.^ For the 

1 W. of Malmesbury, Gesta Pont. loc. cit., p. 255, states that Wilfrid brought 
masons with him from Rome to carry out his English work. The early 
authorities do not tell us this, and for 'Rome' we should certainly understand 
* Romanized lands,' especially Gaul. 


seats of the new experiments in construction and planning 
which transformed the Early Christian into the mediaeval 
style, we should probably be right in looking away from 
Rome altogether to the more northerly districts of Italy, 
the metropolis of which is Milan, and beyond the Alps to 
centres like Cologne, or Tours, or our own York. Here 
classic and barbarian elements of culture met and mingled, 
and though there may not be anything specially Teutonic 
in the forms thus evolved, it may well have been the case 
that the contact with new social forces embodied in the 
northern peoples stimulated inventive genius in the alumni 
of the older schools. The centuries from the sixth to the 
ninth must have been in this respect of great importance, 
and it is a matter of some national pride to find that, in 
one conspicuous instance, these churches built by Wiltrid 
Hexham at the close of the seventh and beginning of the 
eighth century, our country was doing its full share of work 
in the formation of the mediaeval styles. 

For in what Eddius and Prior Richard tell us, the special 
interest lies in the fact that Wilfrid erected in the same place, 
though not at the same time, both a basilican church and one 
of a central type, and appears to have employed for the first 
some of those more advanced architectural forms that belong 
historically to the latter. The buttressing up of a central mass 
with side-buildings so as to form a varied composition, is what 
seems to be implied in the descriptions of the larger church, 
and this is the special characteristic of the central church, 
noted as belonging to it as early as the time of Constantine, 
whose polygonal edifice at Antioch is described by Eusebius^ 
in words that read like a Greek edition of Prior Richard's 
Latin eulogy of the Hexham basilica. The galleries which 
are implied in the mention of the winding staircases of 
stone, are features of the central church, as we find it at 
San Vitale, and the ' three stories ' would be formed by such 

' ^It. Const. Ill, 50. 


galleries between the arcades on the ground floor and the 

The multiplication of altars is a Romanesque feature into the 

ecclesiastical reasons for which we need not enter here. At the 

beginning of the ninth century the church on the Plan of St. 

Gall possessed seventeen altars and we are reminded of the thirty 

altars of the church at York in the century before. Wilfrid's 

Hexham church had ' oratoria quam plurima ... in quibus 

altaria ' . . . but whereas on the St. Gall Plan the altars are 

just located in the nave and aisles without any architectural 

provision for their reception, Wilfrid's were in those adjuncts 

with which he ' encircled on every side the body of the 

church,' and this would imply some architectural provision for 

the altars, like the side-chapels and apses, that became so 

common in Romanesque buildings, but in Early Christian times 

were confined to the central churches. We cannot say that 

Wilfrid directly borrowed these features for his basilican church 

of St. Andrew from his central church of St. Mary, for as a 

matter of fact the latter was posterior in date, but Wilfrid or 

his architectural advisers must, one would think, have had the 

central form of church with its structural possibilities in their 

minds all the time. 

In the architectural work which Wilfrid and his builders were 
doing they were not of course alone. The accounts we possess 
of some of the more important churches in Gaul, of a period 
still earlier than Wilfrid's, give us the same impression of a bold 
and innovating treatment of the traditional Early Christian forms. 
The church built by Bishop Naumatius of Clermont, the 
husband of the art-loving dame,^ in the fifth century, and still 
more the basilica erected by Bishop Perpetuus of Tours over 
the tomb of St. Martin about 470 a.d., deserve the title 
epoch-making, and the latter is credited by some modern 
authorities with originating the noble French tradition of the 
ambulatory round the apse. There seems however fair 

^Vol. 1, p. 127. 


ground for claiming for Wilfrid a certain priority in starting 
that transference of features of plan and construction from the 
more architecturally advanced central buildings to the basilicas, 
on which the development of mediaeval church architecture so 
largely depended. 

The claim here made for Wilfrid's work as bold and inno- 
vating may seem somewhat forced in face of the fact that later 
Saxon work showed no special distinction, and yielded ulti- 
mately without a struggle to that of the Normans. This 
is only however in accordance with the phenomena of Saxon 
history in general, in which, as we have already noticed, seasons 
of brilliant promise are succeeded by long eras of national 
ecHpse. It is from this point of view quite in accordance with 
nataral likelihood that the age of conversion was one of such 
stimulus to the artistic powers of the people that a level of 
effort and achievement was reached which subsequent genera- 
tions were not able to maintain. The carved crosses and the coins 
certainly degenerate in artistic value as the centuries pass away, 
and the fine barbaric gold and encrusted work is early in date. 
So too the architectural efforts of the seventh century may well 
have shown an originality and vigour of which the style was 
never afterwards capable. 

These considerations will bear out what was said at the 
opening of this chapter, and may serve to correct any unduly 
depreciative estimate of Saxon architecture which might be 
formed on a survey of the existing remains alone. Could 
we restore in thought the earlier monuments which have 
perished, our estimate of Saxon buildings might be a higher 

It is probable however that the architectural treatment of 
elevations would have been everywhere the same, and we must 
picture to ourselves the plans just under discussion carried out 
with that mingling of originality and force with clumsiness which 
gives its stamp to all the achievements of the Saxon masons. 
Over and above his innovations in planning, the Saxon 


caementarius had practical skill, as well as ideas as to effect, 
in building, though he is constantly betraying his amateur- 
ishness and want of discipline in the orthodox traditions of 
his craft. He could put his materials together in workman- 
like fashion, for the very thin walls which he inherited from 
the Roman builders have lasted well through the centuries, 
and can bear a considerable superstructure. The walls 
of the tower at Barton-on-Humber, which was increased 
in height in later times, are as thin as those of the 
western adjunct, yet the tower is perfectly solid to this day. 
Monkwearmouth and Bardsey towers rest on walls not 2 ft. 
in thickness. 

The idea of the megalithic was very commonly at work 
in the Saxon builder's mind. He uses big material whenever 
he can procure it. The large squared stones of his quoins, 
his flat lintels, and slabs of large superficial area that line his 
door-jambs, are often but by no means always Roman stones 
re-used. Where these last are not available he cuts the blocks 
for himself, as at Earls Barton and Worth (ante, pp. 187, 236), 
and is not averse from the trouble of hoisting these aloft, 
as for the top of the window openings at Escomb (Fig. 67, 
ante, p. 115). The trapezoidal impost is a good illustration 
of his practice (Figs. 46, 59, ante, pp. 96, 108). 

The Saxon builder possesses as we have seen his own 
stock of forms, and in consequence his work, when any details 
are present, is as a rule easily recognized by its distinction 
from the Norman which succeeded to it. There are it is 
true ' transitional ' buildings about which a decision between 
Saxon and Norman is difScult (ante, pp. 82, 217), but these 
could only be properly discussed in connection with Norman 
architecture generally. A good illustration of the distinction 
which generally obtains between the two kinds of work is given 
by the two blocked doorways shown in Fig. 172 bis. Both look 
equally antique and both are generally reckoned pre-Conquest, 
but the square hood-mould and imposts formed with tile-like 



pieces at A, are just as characteristically Saxon as the joggled 
lintel and tympanum filled with stones set diamond fashion 
of B are characteristically Norman. Finally the Saxon designer 
is beyond question a man of some initiative, a seeker, or per- 
haps only a groper, after architectural effect, and work like 
the enrichment of the wall 
surfaces at Earls Barton and 
Bradford-on-Avon, or on 
the nave at Geddington, is 
carefully schemed though in 
parts quite ungrammatical. 

The architecture thus pro- 
duced had not consistency 
and method enough to con- 
stitute in the technical sense 
a style, but there were in 
it qualities which might have 
been worked out under 
favourable conditions into a 
style. It has been described 
above (ante, p. 69) as con- 
stituting a ' province of 
Austrasian Romanesque,' 
but it was an autonomous province, whose alumni dealt with 
the common stock of forms in independent fashion and held 
with tenacity to certain peculiarities which were their own. 
We may take leave of this curious architectural phase, which 
must always possess for ourselves a high degree of interest, 
with a parting glance at a very characteristic monument with 
which we have already made acquaintance, the Late Saxon 
church at Bosham in Sussex (vol. i, p. 104). 

Bosham presents to us a Saxon western tower, a nave 
either of Saxon fabric or on Saxon lines, a chancel arch 
that is one of the most characteristic specimens of the period, 
and an original chancel greatly extended in later times. 

Fig. 172 bis. — Blocked doorways. 

A. Miserden, Gloucestershire [Saxon]. 

B. Hatfield, Herefordshire [Norman]. 











C -^ 
o .^ 






■- O 









f'^g- ^73 g^ves the plan.^ The tower is quite unadorned and 
its original belfry openings are blocked. The apertures 
toward the nave have been noticed (ante, p. i 70). The plan, 
as", will be seen at a glance, has been set out with more than 
mediaeval indifference to exactness of measurements and squar- 
ing, and the chancel diverges phenomenally from the axis of 

Fig. 174. — Jamb of chancel arch, Bosham. 

the nave. The elevations are gaunt in their plainness and 
the now unplastered rubble work is rough and uncomely, 
but the dimensions are ample, the walls lofty, and the chancel 
arch undeniably imposing. Fig. 174 shows the lower part 
of the northern jamb, and there is no feature in any Saxon 
building that is more characteristic. The jamb, which 

^The writer thanks Mr. Edward S. Prior for kind help in the preparation 
of the plan of Bosham. 


possesses a soffit shaft and angle shafts, is bedded on two 
huge slabs, a square one measuring 4 ft. west to east and 9 in. 
high, and another above it in the form of a circular disc 
3 ft. 6 in. in diameter by 9 in. in height. These slabs are 
commonly attributed to the Romans, but it is not easy to 
see what part of a Roman building they can ever have formed. 
The truth is that they bear no resemblance to known classical 
features, while they are on the other hand characteristically 
Saxon, The nearest parallel to them is to be found in the 
imposts of the chancel arch at Worth, Sussex, a place far 
away from Roman sites. The Worth imposts, like the bases at 
Bosham, are huge and ungainly, testifying both to the general 
love of bigness in the Saxon builder, and his comparative 
ignorance of the normal features which in the eleventh century 
were everywhere else crystallizing into Romanesque. Saxon 
England stood outside the general development of European 
architecture, but the fact gives it none the less of interest in 
our eyes. 



In the following list will be found in alphabetical order the 
names of the places where masonry of Saxon character is 
still to be seen, while the accompanying map, at p. 344. 
indicates their local distribution. The criteria according to 
which a place on the list has been adjudged or refused 
have been sufficiently explained in what has gone before. 
No account has been taken of the mere appearance of 
antiquity in a building, nor of local or historical considerations 
which may point to a pre-Conquest date for particular 
examples. The inclusion of a church in the list has been 
determined by the appearance or definite features which are 
known to be Saxon. These features are in any case worth 
cataloguing, though in a few isolated instances they may 
represent a survival of Saxon forms in post-Conquest build- 
ings. The percentage of such survivals is probably greatest 
in the East Anglian region, where the Saxon peculiarity 
of the double-splayed window appears in what must certainly 
be Norman work on the western side ot the cloisters at 
Norwich cathedral. So far as that region is concerned, the 
fact casts a doubt on the validity of this particular criterion, 
and wherever in that part of England we have only double- 
splayed windows to judge by, some uncertainty must attach 
to decisions. In other parts of the country reliance on 


special features of the kind seems thoroughly to be justified. 
Where they are present other considerations are almost 
always in favour of a Saxon ascription. 

There are numerous other buildings in different parts of 
the country that are possibly in part pre-Conquest, and in 
a few there is a strong probability that this is the case, but 
in the absence of definite indications these have all been 
omitted. The following rejected examples may be mentioned 
to show the class of building referred to : — Stamfordham, 
Northumberland ; Aycliff, Durham ; Hawkeswell, Ainderby 
Steeple, CoUingham, Hooton Pagnell, Maltby, Stainton-by- 
Tickhill, Yorkshire ; Sandiacre, Sawley, Derbyshire ; Off- 
church, Warwickshire ; Rushbury, Shropshire ; Iver, Bucks ; 
Cholsey, St. Leonard Wallingford, Berkshire ; Minster-in- 
Sheppey, Shorne, Leeds, Gheriton, Kent ; Old Shoreham, Ford, 
Lyminster, Sussex ; Netheravon, Wiltshire ; Ashchurch, Up- 
leadon, Gloucestershire ; St. Woolos, Newport, Monmouth- 
shire ; Tintagel, Cornwall. 

In the case of each example included in the list there is 
furnished a very brief indication of the amount of Saxon 
work to be seen in the building, and of any features of 
special interest which this work may offer. It is of course 
impossible in the few words used to give a complete 
inventory, and all that is offered is some general guidance 
as to what the investigator may expect to find upon the 
spot. Where the building or any feature of it has been 
discussed in the text a reference is given to the page. 

The indications as to date are of a general kind, and the 
significance of the letters A, B, C,^ C,^ C,^ has been already 
explained, ante, p. 289 f. The criteria of date are discussed, 
ante, p. 273 ff., and it may be repeated that the ascriptions 
to period B, corresponding to the Danish epoch of 800- 
950, are in most cases merely tentative. Only in the case 
of St. Michael, St. Albans, do we get literary evidence of a 
date within these limits. The middle sub-period of the 


last epoch, C,^ corresponding to the time of Cnut, can 
claim on documentary evidence two examples, Greenstead, 
Essex, and (in part) Stow, Lincolnshire ; the few other 
ascriptions to the time are conjectural. The examples which 
only present long-and-short work, as we have seen (ante, 
p. 289) may belong either to period C or period B. 

It should be explained that the words ' tentative ' and ' con- 
jectural ' do not imply that dates have been assigned by 
guesswork. In each case there is some definite reason for the 
ascription, though the evidence in its favour may not be con- 
clusive. Some critics have adopted a summary method by 
which a few examples have been accepted for the seventh 
century and all the rest congregated in the first half of the 
eleventh. Now it is true that in Europe generally the latter 
period was one very fertile in new architectural undertakings, 
but in England it seems to have been the last half of the tenth, 
the time of Edgar, that showed this special activity. It is a 
significant fact, attested by the episcopal lists given in Stubbs's 
Registrum Sacrum Anglicanum, that certain bishoprics were in 
abeyance during parts of the centuries of Danish ravage, and 
were reconstituted in the time of Edgar. The revival of the 
bishopric meant of course the rebuilding or restoration of the 
episcopal church, and considerable building activity in the 
parishes round about. In the same reign also the Anglo-Saxon 
Chronicle tells us (ad ann. 963) that iEthelwold of Winchester 
begged of King Edgar ' all the minsters which heathen men 
had formerly broken down, because he would restore them : 
and the king cheerfully granted it.' This points to an extensive 
restoration of monasteries for which too we have other evidence, 
while we know that some abbey churches were restored at the 
same time as parochial. Hence a certain amount of English 
work, that might be of the eleventh century, but which shows 
early indications, has in the following list been assigned to the 
last half of the tenth. 

Even in the Danish period we have notices of church 


building, such as that which tells of Swithun of Winchester 
who died in 862, how that he was 'a diligent builder of 
churches in places where there were none before, and a repairer 
of those that had been destroyed or ruined.' ^ Hence when 
indications seem to point to a date betore the Edgar revival for 
a church that does not at the same time seem one of the 
earliest, an ascription to period B may, provisionally, be 

These explanations may redeem from a charge of dogmatism 
the indications of date which follow, and it is believed that the 
reader will be glad of, or even expect, such an expression of 
the writer's views as to chronology, even though it make no 
pretensions to finality. 

'^ Acta Sanct. Jul. i, 291. 


[After name and county there follows an indication of date by means of 
the capital letters A, B, etc. The words within brackets give a summary of 
the amount and character of Saxon work now visible, and references immedi- 
ately following the brackets direct to the page or pages where the general 
notices of the examples will be found. References to special features follow. 
An asterisk implies an illustration.] 

St. Albans, Herts, St. Michael, B (nave of c. 950, arcades later), date, 
288 ; windows, 298. 

Alkborough, Lincolnshire, C^ (western tower). 

Appleton-le-Street, Yorks, C^ (western tower). 

Arlington, Sussex, C^ (main fabric), 292. 

AvEBURY, Wilts, B (main fabric) date, 288 ; windows in two tiers, 172, 298. 

Bardsey, Yorks, B (possible porch-tower), 56; plan, 156*; shaft, 200*; 
thin walls, 326. 

Barham, Suffolk, C or B (fragment of long-and-short work). 

Barholm, Lincolnshire, C^ (enriched south door), 181 * ; string course, 236. 

Barnack, Northants, C^ (western tower and details), 205 f., 185, 208, 214; 
date, 290; doorway, 103; imposts, 181*, 205; pierced mid-wall 
slab, 203*; pilaster strips, 36; plan, 205*; recesses, 207; tower 
arch, 127, 206*. 

Barrow, Salop, C^ (chancel). 

Barton-on-Humber, Lincolnshire, C^ (tower and western adjunct), 208 f. ; 
baluster shafts, 91*, 175, 196*; cap, 179-80*; date, 291 f; door- 
way, 103; double-splayed lights, 93*; mid-wall slab, 204; mould- 
ings at door, 258; plan, 210*; section, 215*; tower, 238; tower 
arch, 212, 291 ; views, 208*, 210*. 


Bedford, St. Peter, C (tower, chancel), 226; date, 291. 
Bessingham, Norfolk, C^ (round western tower), 184-5*. 
BiBURY, Gloucestershire, C (traces). 
BiLLiNGHAM, Durham, C^ (western tower). 

BiRSTALL, Leicestershire, C (window with pierced mid-wall slab in chancel), 

BiSHOPSTONE, Sussex, B or C (south porch, part of nave), i3of., 137, 155, 227; 

plan, 131*; sundial, 131 ; view, 132*. 
BoARHUNT, Hants, C^ (main fabric), 104 f ; date, 106 ; pilaster strip in gable, 

106 ; plan, 105*; view, 106*; western division, 217 ; window, 93*. 
BoLAM, Northumberland, C^ (western tower), corbel cap, 63. 
BoLNEY, Sussex, C (south doorway and perhaps fabric). 
BosHAM, Sussex, C^ (complete, chancel lengthened), 172, 327 f ; chancel, 211, 

280; chancel arch, 109,291,329*; plan, 328*; tower, 157-8, 171 *-2. 
St. Botolph, Sussex, C^ (chancel arch). 

Bracebridge, Lincolnshire, C^ (western tower, main fabric), caps, 177 f., 
180*; squint, 1 28 ; view, 178*. 

Brad ford-on- Avon, Wilts, C ^ (complete, nave, chancel and north porch), i 3 1 f. ; 
arcading, 135*; carved angels, 139 ; chancel arch, 127, 138 ; date, 73, 
287; doorway, 127; an ecclesiola, 79; height of walls, 137; plan, 
132*; porch, 130,227; view, 134*. 

Branston, Lincolnshire, C^ (western tower), arcading, 161*; cap, 177, 180*. 

Breamore, Hants, C^ (complete, large church partly cruciform), 226, 232 f. ; 
inscription, 234-5*; plan, 233*; tower, 239; view, 234*. 

Bremhill, Wilts, C or B (long-and-short quoins). 

Brigstock, Northants, C^ (western tower with stair turret, nave), open- 
ings, 94*, 297; pilasters, 97*, 109; stair turret, 175; tower arch, 97*. 

Britford, Wilts, C^ (nave, enriched archways to side chapels), 226, 227 f. ; 
archways, 229*; date, 288; plan, 22S*; Roman technique, 228 f. ; 
strip-work, 229. 

Brixworth, Northants, A and C^ (large basilican apsidal church with modi- 
fications in later Saxon times), 246 f . ; ambulatory, 250; apse, 247; 
arcade or screen, 249; baluster shafts, 198*, 253; buttresses, 89, 250; 
chancel, 250 f.*; crypt, 250, 267; date, 273 f ; dimensions, 316; plan, 
248*, 295, 316; Roman materials, 246-7*; side aisles, 246; stair 
turret, 247-8, 254; tower, 248, 252 f., 255*, 302 f; view, 246*, 
252*; window, 248. 

Broughton, Lincolnshire, C^ (tower formerly body of the church, stair turret), 
211 f ; arches, 213*, 291 ; cap, 176; turret, 175. 


N. BuRCOMBE, Wilts, C or B (long-and-short quoins to chancel). 
Lt. Bytham, Lincolnshire, C or B (long-and-short fragment). 
Bywell, St, Andrew, Northumberland, C^ (western tower), 161. 

Cabourn, Lincolnshire, C ^ (western tower). 

Cambridge, St. Benet, C- (western tower with details, fine tower arch), finish 
of tower, 163*; tower arch, 291. 

Cambridge, St, Giles, C^ (tower arch preserved in modern church). 

Canterbury, St, Martin, A (western part of chancel and nave), 119 t". ; buttress, 
120; date, 273 f . ; plan, 120*; plastering, 121; proportions, 279; 
side chapel, 121, 227, 236; windows of nave, 122. 

Canterbury, St. Mildred, B or C (south wall of nave and part of chancel), 
quoin, 86*, 277. 

Canterbury, St. Pancras, A (foundations nearly complete of single-celled 
apsidal church with side chapels and porch), 122 f ; altar, 125 j apse, 124; 
arcade or screen, 123, 128] buttresses, 122-4; chapels, 124, 227; date, 
273 f. ; doorway, 122; materials, 124, 258 ; plan, 123*, 294; propor- 
tions, 279; porch, 124, 130, 155 ; W. Thorn on, 125. 

Carlton-in-Lindrick, Notts, C^ (enriched tower arch). 

Caversfield, Oxon, C (western tower), date, 291. 

Clapham, Beds, C^ (western tower), 160*. 

Claydon, Suffolk, C or B (long-and-short quoins to nave). 

Clayton, Sussex, C^ (chancel arch), 109*. 

Clee, Lincolnshire, C^ (western tower), cap, 177, 180*; doorway, 164*; 
keyhole loop, 165*; plinth, 85*; tower arch, 166. 

Clee, St. Margaret, Salop, C^ (herring-bone facing like Diddlebury). 

Colchester, Trinity Church, C (western tower). 

Coleby, Line, C^ (western tower). 

CoLN Rogers, Gloucestershire, C^ (complete nave and chancel, western tower 
later), 109; impost of chancel arch, 181*; proportions, 279. 

CoLTisHALL, Norfolk, C (north wall of nave). 

Corbridge, Northumberland, A (western porch-tower, nave), i 5 i t, ; date, 287 ; 

gable cross, 211*; porch, i 30 ; plan, 151*; portal, 153*; tower finish, 

CoRHAMPTON, Hants, C^ (complete), supposed apse, 281; impost, 181*; 

pilaster strip above doorway, 182. 
Gt. Corringham, Lincolnshire, C^ (western tower), 
Cranwell, Lincolnshire, C or B (part of nave with long-and-short work). 



Daglingworth, Gloucestershire, C (main fabric), impost, i8i*; western 
division, 105, 219. 

Darsham, Suffolk, C or B (north door of nave, long-and-short quoins). 

Debenham, Suffolk, C (western tower). 

Deerhurst, Gloucestershire, C^ or B (complete Saxon monastic church with 
later additions, western tower, nave, transeptal chapels, apse), 299 f,, 
apse, 118; date, 301 ; double opening in tower, 300*; gallery at western 
end, 172 ; plan of eastern end, 231*; of tower, 299*; recesses in tower, 
168; side chapels, 232; tower, 157, 161, 170 f., 171*, 243, 299*. 

Deerhurst Chapel, C^ (complete), 109 f.; impost, 181*, 182; view, 109*. 

Diddlebury, Salop, C^ (north wall of nave), 218 ; herring-bone facing, 51, 89, 
218*; openings, 93*, 95*, 103; plinth, 218 ; staple for door, I03. 

Dover, St, Mary in the Castle, C^ (complete, central towered, cruciform 

church), plan 307* ; tower, 239, 243 ; transepts, 235 ; transeptal arches, 

Dunham Magna, Norfolk, C^ (nave and axial tower), 224 f. ; apse (later), 281 ; 

arcading in interior, 137* ; plan, 225*; plinth, 85*; tower, 239; view, 


Earls Barton, Northants, C^ (fine western tower with details), 184 £, 326-7; 

baluster shaft, 199*; belfry openings, 91, 190*; date, 290; doorway, 

103, 187*; pilaster strips, 89*, 276*; plan, 186*, 190, 214*; string 

course, 181, 276*; view, frontispiece*. 
N. Elmham, Norfolk, C^ (complete plan), 219 f . ; plan, 223*. 
EscoMB, Durham, A (complete nave and chancel church), no f . ; date, 273 f. ; 

doorways, 103, 114*; plan, iio*; proportions, 83, 279; sundial, 115; 

view, III*, do. int., 113*; windows, 1 1 5 *. 

Geddington, Northants, C (arcading on north wall of nave), 277, 327. 

Glentworth, Lincolnshire, C^ (western tower), cap. 180*, 276 ; shaft, 175. 

GoDALMiNG, Surrey (traces now almost obscured), 226. 

GosBECK, Suffolk, C or B (long-and-short quoins to nave). 

Greens Norton, Northants, C or B (nave well preserved, long-and-short 

quoins, no openings). 
Greenstead, Essex, C^ (nave built of wood of c. 1020), 40 f. ; plan, 42*; 

view, 41 *. 
Guildford, C (tower formerly axial), 225 ; date, 291. 

Hackness, Yorks, B or C (chancel arch with carved impost). 
Hadstock, Essex, C^ (nave, double-splayed windows). 


Hainton, Lincolnshire, C^ (western tower), plinth, 85*. 

Gt. Hale, Lincolnshire, C^ (western tower), 161 ; caps, 180* ; plan with turret 
stair, 175*. 

Gt. Hallingbury, Essex, C^ (chancel arch). 

W. Hampnett, Sussex, C^ (chancel), narrow loop, 93*, 273, 296. 

Harmston, Lincolnshire, C^ (western tower). 

Harpswell, Lincolnshire, C^ (western tower), 159. 

Hart, Durham, B (fabric of nave), baluster shafts, 195. 

Headbourn Worthy, Hants, C (main fabric of nave and chancel). 

Heapham, Lincolnshire, C^ (western tower). 

Heddon-on-the Wall, Northumberland, C or B (long-and-short quoin). 

Hemingstone, Suffolk, C or B (long-and-short quoin). 

Herringfleet, Suffolk, C^ (round western tower), strip-work, 184. 

Hexham, Northumberland, A (crypt) 264-5 *f.; date, 273, 317; Roman 
stone at, 192*. 

[Hexham, Wilfrid's churches at, see General Index.] 

Heysham, Lancashire, C (western door and old north door), chancel arch, 105. 
Heysham Chapel, B (ruins of single-celled chapel), 30, 79; date, 279; door- 
way, 102*; plan, loi*; proportions, 279. 

HiNTON Ampner, Hants (traces of Saxon work). 

Holton-le-Clay, Lincolnshire, C^ (western tower). 

Hough-on-the Hill, Lincolnshire, C (western tower and stair turret), 175; 

date, 291. 
Houghton-on-the Hill, Norfolk, C (fabric of nave), mid-wall slab, 204. 
HoviNGHAM, Yorks, C^ (western tower). 
Howe, Norfolk, C (round western tower), 93*- 

Jarrow, Durham, A (present chancel, baluster shafts in porch), 140, 194*, 
195 ; carved stone, 193*; corbel cap, 64*; date, 273 ; tower, 150. (See 
also General Index). 

Kingsbury, Middlesex, C or B (long-and-short western quoins). 
Kirk Hammerton, Yorks, C^ (complete Saxon fabric with later additions), 
doorway, 98*, 165. 

KiRKBY Hill, Yorks, C^ (parts of fabric). 

KiRKDALE, Yorks, C^ (nave, sundial over south door), doorway, 98*, 165; 
pilaster strip in gable, 106. 


Langford, Oxon, C^ (axial tower, details), 225 ; keyhole aperture, 166. 

Laughton-en-le-Morthen, Yorks, C (north door of nave). 

Ledsham, Yorks, C^ (main fabric late Saxon) tower, 156. 

Leicester, St. Nicholas, C (nave). 

Lewes, St. John sub Castro, C (doorway preserved). 

E. Lexham, Norfolk, C (round western tower), mid-wall slab, 188. 

Lincoln, St. Benedict, C^ (western tower). 

Lincoln, St. Mary le Wigford, C^ (western tower), cap, 176, 180*. 

Lincoln, St. Peter at Gowts, C^ (western tower, west quoins of nave), 
160*, 171*, 178; cap, 179, 180*. 

Lindisfarne, C (remains of Saxon chancel and apse), apse, 118. 

Lydd, Kent, C^ or B (remains of basilican church), 24.5 ; plan, 245*. 

Lyminge, Kent, A (foundations of apsidal church), 118; arcade or screen, 128 ; 
date, 273. 

Market Overton, Rutland, C (tower arch preserved), 96* ; impost, 108, 181. 

Marton, Lincolnshire, C^ (western tower). 

W. Mersea, Essex, C (western tower). 

Middleton by Pickering, Yorks, C^ (western tower), 156. 

Miserden, Gloucestershire, C (north and south doorways), 327*, impost, 

Monk Fryston, Yorks, C^ (western tower). 

Monkwearmouth, Durham, A (porch-tower and nave), 140 f ; western arch- 
way, 143*; baluster shaft, 145*; carving on jamb, 144*; date, 74, 148 f., 
273, 295; doorways, 103; plan, 141*; porch, 130 f., 155, 303; pro- 
portions, 83, 279; thin walls, 326; vault, 127, 143, 169; view, 142*; 
west wall, 147*, 170; windows, 93*. 

Nettleton, Lincolnshire, C^ (western tower). 

Newton, Norfolk, C^ (axial tower), 225. 

Northleigh, Oxon, C^ (axial tower), 225. 

Norton, Durham, C^ (cruciform, central tower), 306 f. ; transepts, 235; 
transeptal arches, 235; tower, 239. 

Norwich, St. Julian, C^ (round western tower). 

Ovingham, Northumberland, C^ (western tower). 

Oxford, Cathedral, B or C (indications of triple apses), 244; date, 288. 

Oxford, St. Michael, C^ (western tower), 161. 


Pattishall, Northants, C^ (main fabric), 77; impost, 181*. 

[St. Peter-on-the-Wall, Essex, see pp. 116 f., 289.] 

Peterborough, A (foundations of Saxon church), 227, 316; date, 287; plan, 
240, 286, 315*; presbytery, 240; transepts, 240. 

Reculver, Kent, A (foundations of part of basilican church ; columns at Canter- 
bury), 255 f. ; apse, 118; arcade or screen, 128; columns, 256, 259'*; 
date, 273; plan, 256*. 

Repton, Derbyshire, C^ (chancel and eastern end of nave, crypt), 232; 
columns, 232, 258-9*; crypt, 269* f. ; date, 292 ; side chapels, 232 ; plan, 
231*; tower (problematical), 238. 

RipoN, A (crypt), 264-5* f- ; '^ate, 273, 317. 

Rochester, A (foundations of apsidal church), 119; arcade or screen, 128; 

a bishop's church, 309; date, 273 f; plan, 119*, 294; proportions, 


Rockland, Norfolk, C^ (nave, long-and-short quoins with late indications). 

RopsLEY, Lincolnshire, C or B (nave, long-and-short quoins). 

Rothwell, Lincolnshire, C^ (western tower, south-west quoin of nave), 177; 
cap, 177, 180*. 

Scartho, Lincolnshire, C^ (western tower), cap, 179, 180*, 276. 

ScoLE, Norfolk, C or B (long-and-short fragments). 

Sherborne, Dorset, C (remains of western door), a bishop's church, 309. 

Sidbury, Devon, C^ (crypt), 270-1*. 

Singleton, Sussex, C (western tower with double-splayed lights). 

Skillington, Lincolnshire, C or B (long-and-short fragments). 

Skipwith, Yorks, C (western tower, recess), 168*. 

SocKBURN, Durham, B (remains of fabric of nave). 

Lt. Sombourn, Hants, C (nave, pilaster strips). 

SoMERFORD Keynes, Wilts, B or C (north door of nave), 75, 102*; date, 279. 

SoMPTiNG, Sussex, C^ (western tower, tower arch, caps), 200 f ; corbel cap, 64*, 

202*; tower arch, 201*, 291 ; tower finish, 163*, 201 ; German 

influence at, 276. 
Springthorpe, Lincolnshire, C^ (western tower). 
Stanton Lacy, Salop, C^ or ^ (nave, pilaster strips, transept), 226, 235 ; pilaster 

strip, 182, 278. 
Stevington, Beds, C (western tower, details), mid-wall slab, 204; date, 291. 
Stoke d' Abernon, Surrey, B or C (part of north wall of nave ; sun-dial). 


Stone-by-Faversham, a (western part of chancel and eastern end of nave), 
masonry of Roman type, 287*. 

Stopham, Sussex, C^ (north and south doors), mouldings, 98*. 

Stow, Lincolnshire, C^ and - (transepts, tower arches), 235, 292 ; a collegiate 

church, 309; date, 292; plinth, 85*, 293; quoin of transept, 87*; 

tower, 239-40*. 

Stowe-nine-Churches, Northants, C^ (tower arch). 

Stragglethorpk, Lincolnshire, C^ (western door, similar to that at Dunham 

Swallow, Lincolnshire, C^ (western tower). 
SwANSCOMBE, Kent, C (western tower), date, 291. 

Thurlby, Lincolnshire, C (western tower). 
TicHBORNE, Hants, C^ (chancel). 

Warblington, Hants, C (tower), 174. 

Warden, Northumberland, C^ (western tower), imposts of Roman worked stones, 

Wareham, St. Martin, Dorset, C^ (nave and chancel), 106; double open- 
ing, 158; plinth, 85*; proportions, 279; window, 93*. 

Waith, Lincolnshire, C^ (axial tower), 225. 

Weybourn, Norfolk, C (tower originally over chancel), 226. 

Wharram-le-Street, Yorks, C^ (western tower), 156. 

Whitfield, Kent, C^ (nave and chancel with additions), plan, 104*. 

Whittingham, Northumberland, C^ (western tower). 

WicKHAM, Berks, C^ (western tower), Roman shaft in belfry, 10*, 194. 

WiLSFORD, Lincolnshire, C or B (long-and-short quoin to nave). 

Wing, Bucks, C^ (complete basilican church with crypt), 259 f. ; apse, 118, 
259 f. ; arcading, 276; baluster shaft, 199*; crypt, 268*; date, 73; 
double opening, 158; plan, 268*; view, 260*. 

WiNTERTON, Lincolnshire, C^ (western tower). 

Wittering, Northants, C^ (nave and chancel, with additions), 107 f. ; 

chancel arch, 108*, 291; impost, 181 ; long-and-short quoin, 88*; 

plan, 107*. 
WiTTON, Norfolk, C^ (round western tower, north side of nave), plan, 184*. 
WooLBEDiNG, Susscx, C (pilaster strips in nave), 90*. 
WooTTON Wawen, Warwickshire, C (tower with arched openings on the 

four faces), 226, 239 f ; transeptal arches, 240. 


WoRLABY, Lincolnshire, C^ (western tower). 

Worth, Sussex, C^ or - (complete apsidal, towerless, cruciform church), 236 f. ; 
apse, 118; chancel arch, 127, 236; doorways, 236; double opening and 
baluster shaft, 92*, 158, 200; imposts, 236, 330; pilaster strips, 236; 
plan, 237*; stringcourse, 236; transepts, 236; transeptal arches, 235; 
view, 238*. 

Wroxeter, Salop, B or C (part of north wall of nave), string course, 236. 
York, St. Mary Bishophill Junior, C^ (western tower, recessed tower arch). 


(For existing Saxon remains see previous Index List. An asterisk indicates 

an illustration) 

Aachen, 46 ; minster 31,48, 51,* 52,* 

58, 285, 301. 
Aber, foundations of early chapel at, 

Abernethy, round tower at, 79. 
Abingdon, Saxon monastery at, 263. 
Abinger, Early Norman church at, 

Acta Sanctorum, quoted, 334. 
Adamy, referred to, 50, 59, 295-6. 
Adler, Professor, referred to, 66, 68. 
Albert, Archbishop of York, 257. 
.^thelberht, 125, 130. 
iEthelred, law of, 79. 
iEthelwulf, 43, 46. 
Africa, North and its churches, 1 2, 

13, 15 f, 262, 281, 295. (See 

also 'Algeria.') 
Aisles, choir, 232 ; side, 11, 15, 244 f., 

Alae, 129, 261. 
Alban, St., Abbey church of, 312 ; 

balusters in, 197.* 
Albury, Early Norman church at, 76. 
Altar, 12, 13, 23, 27, 31, 125, 287 ; 

Roman, 191.* 
Alcuin, 34, 46, 319 f. 
Aldhelm, 132, 314. 
Alexander, bishop of Lincoln, 292. 
Alfred, 43. 

Algeria, 1 3 . (See also * Africa, North.') 
Ambrose, St., 282, 
Amplitude of Early Saxon churches, 

Andred, forest of, 236. 
Andrews, St., St. Regulus Chapel at, 


Angels, carved, at Bradford-on-Avon, 

Apertures, subdivided, 61 f., 64,* 
300,* and passim. 

Apollinare, San, in Classe, Ravenna, 
64,* 183, 263. 

Apostles Church, at Cologne, 57. 

Apse, 11-15, 89, 1 16, 1 19, 124, 126-7, 
261, 280; polygonal, 118, 250, 
259; triple, 244; western, at 
Abingdon, 263, at Canterbury, 
260* f. 

Arcade ; as screen before the apse, 
127-8, 249, 255-6,* 294 ; wall, 
134-5,* 161, 188, 260, 276* f. 

Arch; chancel, 28-9,* 96,* 107-8,* 
I27» 236, 329*; Irish, 18, 27f., 
31 ; recessed, 8, 98, 217, 291, 
293 ; Roman, 3, 155; Saxon, i 27, 
247 ; straight-sided, 93 ; tower, 
96,* 97,* 127, i66f., 170*-!,* 
201,* 206,* 217. 

Archaeologia, referred to, 7, 9, 12, 38, 
1 10, 1 16, 1 18, 182. 

Archaeologia Cantiana, referred to, 1 1 9, 
122, 255-6, 287. 

Archaeological Association, Journal, re- 
ferred to, 182, 207, 270. 

Archaeological Institute, * Norwich ' 
volume, referred to, 224 ; 'Win- 
chester' volume, referred to, 245, 
257»,267, 303, 320. 

Archaeological Journal, referred to, ix, 8, 

9' 95» i73> J82, 195, 212, 220, 

243-4:5, 257; 
Archaeological Society ; of Derbyshire, 

270; of Durham, etc., 147. 


. 1 

a t 

• ■ 


? o 
J z 


>■ t 

a 05 ••: 

.-• o 

' o 

i • ' 






5 •^ 

■.. iC 


■^ h- 



• K 

X '■■■ 

■ -r 


Z O 



Ui > 

u E 



























§iVX(Oi^ CimMftCE* 


^ — y eniTroRcT mintonampmcr* pootBEOirit 

^— ^ L* Ml not' 

Fif.. I7S. — Map thowing ihc duiribucion of Chuwho ih»l tihibit rcmaint of Sjion building. 



Archbishop's Chapel at Ravenna, 283. 

Architectural, etc.. Society of Dur- 
ham, etc., 147. 

Arcosolia, 251, 270. 

Ashlar work, 218. 

Associated Societies Reports, 250, 292. 

Astragal, Roman, ic)i*-z* 

Atkinson, Rev. G., referred to, 292. 

Atrium, 14, 15. 

Augustine of Canterbury, 125, 261. 

Augustine, St., monastery of, 130. 

Austrasia, 46 ff., 275. 

Austrasian Romanesque, 49 ff., 263, 

Baluster shafts, Roman, 9, 10,* 191,* 
193 f., 200; Saxon, 91 *-2,* 144- 
5*-7>* 190 ff"., i92*-3*-6*-7*- 
8*-9,* 200,* 290-1, 295, 303. 

Baptistry, 12, 234, 260 f, 303. 

Bases, 8 f , 176, 179, 180,* 258-9.* 

Basilica, forensic, 8 ; Christian, 1 1* ff., 
65, 245 ff., 281, 294, 320 f. 

Batter, of walls, 1 12, 160. 

Bayeux Tapestry, 41, 71. 

Bede, quoted, 294 ; referred to, 2, 37, 
45, 119, 130, 261, 274, 282, 

' Bee-hive' cell, 19 f. 

Belfry, 44, 53, 61, 63 ; openings, 83, 

90, 91,* 158, 175,* 190,* 209, 

287, 290. 
Benedict Biscop, 45, 140, 148. 
Birrens, Roman altar from, 191.* 
Black Death, 310, 
Blandinium, 243. 
Block house technique, 37. 
Bloxam, referred to, viii, 80. 
Boniface, 43, 46. 
Boothby Pagnell, Lincolnshire, 1 59, 

176, 287. 
Bradford-on-Avon, town of, 133. 
Brick, Roman, 3, 6, 8, 11, 49, 117, 

120, 124, 228, 246-7,* 287.* 
Bridge, Rochester, 7. 
Bridges, Roman, 7. 
Builder, The, referred to, ix, 26, loi, 

219, 257.^ 
Bulletin Archiologtque, referred to, 

Burwash, Sussex, 63. 

Butterworth, Rev. G., referred to, 301. 

Buttress, 57f., 89 f., 117, 120, 121, 
122, 124, 250. 

Byzantine origin of cubical cap, sup- 
posed, 68. 

Cable pattern, 105. 

Campanili, 44, 63. 

Campanology, story about, 53. 

Canterbury Cathedral ; Saxon, 1 30, 
157^227, 257, 260* f., 267, 294, 
312, 314; later, 126, 269, 

Cap, capital ; corbel, 62*-3-4,* 180,* 
202*; cubical, 68, 176 f., 180,* 
276; cushion, 68 ; German, 59, 
60, 61 f. ; at Repton and Re- 
culver, 258-9* ; Roman, 8 f., 
10, 6"] ; scalloped, 177, 180*; 
volute, I 'j^ f,, I 80,* 

Capella, 79. 

Cashel, 20, 27. 

Castor, St., Coblenz, 55, 60. 

Catacombs, cubicula in, 34. 

Caumont, de, referred to, 56. 

Cedd, 1 16. 

Cell, bee-hive, 19 f. 

Celtic features, 18 ff., 296. 

Chambers over chancels, 171, 215. 

Chamfer, hollow, 181, 186, 302 ; 
quirked, 181. 

Chapel, 121, 124 f, 129 f. ; side or 
lateral, 227, 232 f, 294 ; tran- 
septal, 302, 304. 

Charles the Simple, 46. 

Chester, 9. 

Chesters, 4,* 5,* 200. See also 'Cil- 

Chevet rectangulairc, 280. 

Chichester, 243, 264. 

Chinon, St. Mexme at, 198. 

Chollerford, 7. 

Chollerton, 8, 259.* 

Chronicle, Anglo-Saxon, referred to, 

Chronicle of Abingdon^ referred to, 

Chronicle of Lanercost, quoted, 312. 

Chronicle of Ramsey, quoted, 241 ; 
referred to, 245, 298. 

Chronological tests, 272 



Churches ; Apostles, 55, 57, 282 ; 
bishops', 119, 309; collegiate, 
309; cruciform, 124, 226 fF., 
282 fF., 321 ; Irish, 18 fF. ; nave 
and chancel, 27-8* and passim ; 
polygonal, 3 2 1 ; Romanesque, 
321 ; Romano-British, 2, 261, 
294; Saxon, passim. 

Cilurnum (Chesters) 4,* 5,* 7, 8, 93.* 

Clara Thurm, at Cologne, 50. 

Clemen, Paul, referred to, 56, 224. 

Clermont, 324. 

Cluniac reform, 242. 

Cnut, 34, 43, 243, 290. 

Coblenz, 55, 60. 

Coenaculum, at Sellgenstadt, 303. 

Cologne, 50, 55, 57, 66, 89, 323. 

Columnar edifices, Roman, 8. 

Columns, 8, 65, 232, 245, 256, 

Compton, Surrey, 171. 

Concrete, 3, 5, 7. 

Confessio, 251. 

Corbel cap, see ' Cap, corbel.' 

Corbelling out, 62. 

Corbels, corbel stones, 31,97,278, 302. 

Core and facing, 49 f. 

Corfe Castle, 71. 

Cormac's Chapel, 27, 29. 

Cornish oratories, 100. 

Cornwall, 18. 

Corvey, 55, 67. 

Cravant, 50. 

Crescimbeni, referred to, 16. 

Criteria, 81 f, 272 fF., 33 i. 

Cross; ornamental, 31,* 77; gable, 
211*; Greek, 130, 282, 286, 
319 ; Latin, 284, 286 ; symbol 
of, in ground plan, 242. 

Crucified, the, 139, 216. 

Cruciform, see 'Churches, Cruci- 

Crux commissa, 221, 241, 283, 286. 

Crux immissa, 284. 

Crypts, 33, 263 fF. 

Cuthbert ; of Canterbury, 261 ; of 
Durham, 311 ; of Jarrow, 45. 

Danish inroads, their influence on 
pre-Conquest architecture, 36 fF., 
297 f. 

Darenth, Kent, 171, 273. 

Defence, as a motive for towers, 56, 
166, 304 f 

Dehio und von Bezold, quoted, 286 ; 
referred to, 14, 56, 58, 68, 280, 
284, 295, 305. 

Denis, St., 283.* 

Derbyshire Archaeological Society, 

Diaconicon, 16. 

Dietrichson, Professor, referred to, 39 f. 

Domesday ; evidence from for Saxon 
churches, 75 f; quoted, 220. 

Doorwa}'s ; in upper stages of towers, 
169 f, 214 f.; Irish, 20,* 23,30*; 
rebated, 103, 230; Roman, 4,* 
5,*8, 152-3* ; Saxon, 94-5,* 98,* 
102,* 114,* 122, 164,* 187*, 
square-headed, 93, 145, 302, 
their narrowness, 127, 236, 297, 
triangular-headed, 93, 224, 226, 
302, and passim. 

Double-splayed windows, see * Win- 

Dover; Roman Pharos at, 3, 122; 
baluster shafts in museum at, 196. 

Dowker, G., referred to, 256. 

Dowth, 19. 

Dryden, Sir H., referred to, 250. 

Ducange, referred to, 129. 

Dulane, 25.* 

Dun ^ngus, 20. 

Dunraven, Lord, referred to, 30. 

Dunstan, 129. 

Durham, 127, 312, 314; Architec- 
tural, etc.. Society, 147. 

Earthworks; Norman, 71, 224; Saxon, 

Eastern ends, flat, 280. 
Eaton Bishop, Herefordshire, 63. 
Ecclesiae quadratae, 21 f. 
Ecc/esio/ogisf, The, referred to, 147. 
Ecgberht, 46. 
Eddius, quoted, 264, 318. 
Edmer, referred to, 314. 
EfFmann, quoted, 53; referred to, 61, 

130, 212, 306. 
Eginhard, 66, 303. 
Elmham, 2 1 9 f ; South Elmham, 2 1 9 f , 

222*; Thomas of, referred to, 274. 



Ely, 241, 312-3. 
Enceintes, town, 70. 
Encorbelment, I9f. 
Enlart, Manuel (C Archeoh^e, quoted, 
280; referred to, 49, 283, 285, 

Entasis, in German columns, 65. 

Essen, 57, 61, 6"]. 

Eusebius, referred to, 323. 

Exeter, 82, 243,* 312-3. 

Fa9ade, Romanesque, 53-4-5 ; twin- 
towered, 55-6, 243, 285. 
Finial, Irish, 3 i. 
Fischbeck, Westphalia, 276. 
Fonts, Saxon, 77. 
Forebuildings, western, 169. 
Framed timber work, 37 f. 
Framingham Pigot, Norfolk, 78. 
Freeman, Professor, referred to, 95, 

Fulda, 66. 

Fursa, 1 1 6. 

Gables, 24, 85, 112, 141. 

Gall, St., Plan of, 183, 283,* 324. 

Galla Placidia, her tomb at Ravenna, 

Gallerus, 23, 24.* 

Galleries, dwarf, 65 ; western, in 
Saxon churches, 172. 

Gateways, see ' Doorway's.* 

Gelnhausen, 276. 

Gentleman's Magazine, referred to, 145, 

Germany, its connection with Eng- 
land, 44, 47, 69, 298 ; its archi- 
tecture, 49 ff., 198, 298. 

Germiny des Pres, 57. 

Gernrode, 48, 55, 60, 276.* 

Gesta Poniifcum, see * William of 

Giovanni, San, a Porta Latina, church 
of, at Rome, 1 6. 

Glass, workers in, sent for from abroad, 
45, 148. 

Glastonbury, 129. 

Glendalough, 28,* 29,* 30.* 

Gloucester, 269, 312. 

Godiva, 292. 

Goodmanham, 212, 

Green, J. R., referred to, 46. 

Gsell, M. S., referred to, 12, 15, 17, 

Guesseria, N. Africa, 17.* 
Gwythian, Cornwall, oratory at, 127. 

Hagioscope, 268, 270. 

Hallenkrypta, 269. 

Handbuch der Architectur, referred to, 

Handle-stone, 3 1 . 
Harold, 43, 71. 

Hartshorne, Albert, referred to, 45. 
Hatfield, Herefordshire, 327.^ 
Height of walls, in Saxon buildings, 

83» 137, HI' 151. 279- 
Helm, the German, 162, 201. 

Henry the Fowler, 47. 

Hereford, 312. 

Herring-bone work, 51, 71,* 85, 218,* 

Hersfeld, 284.* 
Hexham; Wilfrid's churches at, 192,* 

257, 286, 312, 317 f; Richard 

of, quoted, 3 17 f 
Hooton Pagnell, Yorks, 217, 220. 
Hornby, Yorks, 287. 
Housesteads, 10.* 
Howth, near Dublin, 80. 
Hoxne, Suffolk, 220. 
HughWhite (Hugo Candidus), referred 

to, 247, 274.. 
Humann, on cubical caps, 68. 
Hutchinson, History of Durham, referred 

to, 83. 

Imposts, 97, 107, 108,* 1 12, 138, 144, 
167, 180-1,* 199,* 201,* 206,* 
236, 327.* 

Inscription, Saxon, at Brcamore, 

Ireland, 18 ft". 

Irish Romanesque, 32. 

Irvine, J. T., referred to, ix, 36, 

134-5-6-7, 188, 208, 270. 

Iver, Bucks, 95.* 

Jambs, baluster shafts in, 146-7* ; in- 
clination of, 93, 94,* 113, 138, 
I 56, 296 ; masonr)- of, 96,* 98,* 
i02*-3, 1 12-3. 



J arrow ; corbel cap at, 63-4* ; mon- 
astery, 140 ; Norman straight- 
sided arch at, 93 ; tower, 150 ; 
balusters, 193. (See also 'Jarrow' 
in Index list.) 

Jones, Rev. Canon, 152. 

Jouarre, 269. 

Journal of Archaeological Association, 
referred to, 182, 207, 270. 

Jumieges, Early Norman caps at, 68. 

Kapella, Bartholomaus, at Paderborn, 

66 f. 
Keyhole loop, i65*-6, 225. 
Killiney, 28,* 93,* 
Kilmalkedar, 23, 29, 31. 
Kilpeck, Herefordshire, 302. 
Kirchliche Baukunst, see * Dehio u. von 

Kraus, referred to, 15. 

Lanchester, 8, 192. 

Lanfranc, 31 1-2. 

Lastingham, 78, 179, 269. 

Lathe, used for baluster shafts, 9, 144, 

Leathly, Yorks, 166, 304. 

Ledbury, Herefordshire, 93."* 

Leicester, Roman shafts at, 9. 

Length, in plans of Saxon churches, 
83, loi, no. 

Leofric, 292. 

Lewin, Th., referred to, 116, 11 8. 

Lights, see ' Windows.' 

Lincoln, Roman remains at, 7, 8. 

Lincolnshire, its richness in pre-Con- 
quest work, I 58. 

Lindisfarne, Norman balusters at, 198. 

Lion-sur-mer, 159, 162. 

Lisenen, 58 f, 65, 89, 275. 

Long-and-short-work, 7, 82, 87-8,* 
112,273,277, 289,297; in Ger- 
many, 89. 

Loops, see ' Windows.' 

Lorenzo, San, at Milan, 22, 285, 321. 

Lorsch, 50, 58-9,* 67, 301. 

Liibke, referred to, 276. 

Luc-sur-mer, 159. 

Lul, Archbishop of Mainz, 45. 

Macdara, St., oratory, 24-5.* 

Mainz, 9, 60. 

Mailing, West, church of, 89.* 

Malmesbury, Saxon church at, 314; 

William of, see 'William of 

Manor house, 71. 
Masonry, Saxon, 296, 326. 
Maestricht, 54. 
Matthias, St., Trier, 34. 
Mauritius, St.; at Cologne, 55 ; by 

Hildesheim, 259. 
Maursmiinster, 55. 
Megalithic character in Irish buildings, 

29 ; in Saxon buildings, 296, 

Michelstadt, basilica at, 51, 66, 295, 

Mid-wall slab, see ' Slab.' 
Migne, Patrologia, referred to, 45, 303. 
Minden, 57. 

Miniato, San, Florence, 271. 
Moated mound, 42, 71, 224. 
Monasticon, quoted, 40, 292. 
Montmajour, near Aries, 268. 
Mortar ; Irish, 28 ; Roman, 5, 122 ; 

Saxon, pink, 122, white, 122, 

1 24, yellow, 1 24. 
Mortising, 4-5,* 114,* 229*. 
Mosaic, 10, 13, 49, 59*. 
Mos Scottorum, 37. 
Mouldings ; roll, 98,* 109, 144, 291 ; 

cable, 105, 145; hood, 160, 

302; base, 86, 182. (See also 

Much Wenlock, 1 1 8. 
' Mushroom ' cap, 67*-8. 
Mycenae, 19. 

Nantes, 285. 

Narthex, 1 1, 15-6-7. 

Nazaro Grande, San, Milan, 282.* 

Netheravon, Wilts, 174.* 

Neustria, 46. 

Newgrange, 19. 

Newhaven, Norman balusters at, 198. 

Newport arch, Lincoln, 7. ■ 

Niches, 207, see also 'Recesses.' 

Nicholas, St., Caen, 95. 

Niederzell, Reichenau, 66. 

Ninian, 294. 

Nocera, Baptistry at, 62. 



Norman architecture, 38, 48, 56, 57, 

63, 66, 69, 85, 217, 302, 311 f., 

Northmen, 36. (See also ' Danish 

Northumbria, 47. 

Norway, its timber architecture, 39. 
Norwich, 82, 219, 331; volume of 

Archaeological Institute, 224. 
Notre Dame d'Esquay, 56. 

Odda, Earl, 1 10. 

Odo, Archbishop ofCanterbury,26 1-2. 

Ofta, 46. 

'Old Minster' at South Elmham, 

219 f., 222.* 
Oliver, on bishops of Exeter, referred 

to, 243. 
Opus, quadratum, 3, 7; testaceum, 3 ; 

signinum, 7, 121, 255. 
Orientation, 22. 
Ostiarius, 167. 

Oswald, bishop of Worcester, 242. 
Othona, 1 16. 
Otmarsheim, 57. 
Otto the Great, 46. 
Overbury, Worcestershire, 93.* 

Paderborn, 54, 57, 66. 

Pantaleon, St., 55, 66, 89. 

Paris, Thermae at, 8. 

Parker, Sir J. H., referred to, 134. 

Parma, 65. 

Patrick, St., 21, 30. 

Pebble flooring, at Escomb, 115. 

Penmon Priory, 80. 

Periods, of Saxon architecture, 35, 

287 f.; sub-periods, 290. 
Perpendicular clearstories, 83. 
Pertz, Mon. Germ. Hist., referred to, 

Peter-on-the-Wall, St., 89, 1 16 f., 118, 

Peter, St.; at Oxford, crypt, 269 ; at 

Rome, crypt, 263. 
Petit appareil, 49. 

Pharos, Roman, at Dover, 3, 122, 243. 
Pilae, of hypocausts, 9. 
Pilaster strips, Saxon, 42, 58, 74, 89,* 

90,*97,* 106 f., I 34-5,* 182,229, 

236, 273, 275-6,* 298. 

Pilasters; fluted, 59,61, 300;* German, 

58 f., 276,* see also ' Lisenen ' ; 

Irish, 25,* 29 ; reeded, 134, 300. 
' Pilz ' capital, 68. 
Piran, St., oratory of, 100. 
Pisa, 65. 

Planning, Saxon, 316. 
Plastering, Irish, 29, 31; Roman, 7 ; 

Saxon, 88, 115, 121, 255, 266. 
Plinths, Roman, 87; Saxon, 85,* 134, 

161, 218, 273, 293. 
Poitiers, St. Jean at, 49. 
Porch, 28, 117, 124 f., I29f, 151; 

lateral, i3i*-2, 227; western, 

294, 299. 
Porticus, izgf, 242, 281 ; ingressus, 

129, 281. 
Priestholm, 80, 162-3.* 
Proportions, of Saxon plans, 83-4,* 

278 f. 
Prothesis, 16. 

Pudentiana, Sta., corbel cap at, 64.* 
Puffin Island, 80, 162. 

Quattro Coronati, SS., 263. 

Quedlinburg, 68, 276. 

Quoining ; Irish, 29 ; Roman, 7, 1 1 ; 
Saxon, with big stones, 86,* 
87 * f., 112, with long-and-short 
work, 7, 87 f., 88.* 

Ramsey, 241-2, 245, 298. 

Ravenna, 12, 322. 

Rebates, in Saxon openings, 103, 106, 

Ii3» H3. 230- 
Recesses, 173, 207, 299. 
Recessing, 6t,, 64,* 217, 291. 
Regulus, St., chapel at St. Andrews, 

Reichenau, 66. 

Reliquary, T^^, quoted, i 52 f. ; referred 
to, viii., 147, 195, 235, 239, 298. 

Remigius, bishop of Lincoln, 292. 

Repertorium fUr Kumtwissetischaft, re- 
ferred to, 173. 

Restcnnet Priory, 79. 

Rhocn, C, referred to, 5 1 f. 

Richard of Hexham, quoted, 318 f. 

'Roman,' use of the term, 13, 33. 

Romanesque, Austrasian, 49, 69 ; 
Neustrian, 49. 



Romano-British churches, ii f.,* 121, 
260,* 262. 

Rome, its fascination, 43 ; its archi- 
tecture, 322. 

Roofs; stone, 26*; saddle-back, 210 ; 
tower, 162 f * 

Rougemont, Exeter, 82. 

Ruprich-Robert, quoted, 56; referred 
to, 38-9. 

Rusguniae, in North Africa, I 3. 

Sacristan, 167. 

Saint Jean, Poitiers, 49. 

Savenieres, 50. 

Saxony, Old, 47. 

Scandinavian influence on Saxon art, 
36 f , 302. 

Schola, 14. 

Scotland, its architecture, 18, 79, 127. 

Screen, or arcade, 123, 127, 249, 255, 

Sculpture, Saxon ; animal, 291 ; figure, 
138, 146; tectonic, on caps, 
176 ff, 180*; imposts, 181*; 
panels, 230. 

Seligenstadt, 173, 303. 

Selsey, Saxon church at, 243. 

Sens, 34. 

Shafts, 259*; angle, 98, 212, 291 ; 
baluster, 144-5,* 190 (.,* 290-1, 
295 ; mid-wall, 9 1 ,* 1 7 5 f , 1 80,* 
202* ; Roman, 8, 10*, lathe 
turned, 9, in Saxon churches, 
10,* 128,256 f, in later churches, 
8 f, 259* ; soffit, 108 f., 20I*-2, 

Sheppard, Dr., referred to, 257. 

Ship-building technique, 38. 

Shrines, a source of Church revenue, 

Shropshire, Saxon churches in, 80. 
Shutters, in windows, 106, 115. 
Sicily, 56. 

Sidi Mabrouk, Algeria, i 3. 
Silchester, 7, 9, 11,* 16, 38, 245, 294, 
Simeon of Durham, referred to, 150, 

Skellig Michael, 19 f, 22,* loi.* 

Slab, mid-wall, 105,* 203,* 204, 


Soest, St. Patroclus at, 54. 

Square ends, to churches, 27 f , 100 ff., 

Squints, 128, 170,* 173. 
Stainton-by-Tickhill, Yorks, 96.* 
Staircases, spiral, 175, 211, 221, 247, 

318, 321. 
Staple, Kent, 178. 
Staples, door, 103 ; window, 105. 
Statue, at Monkwearmouth, 146. 
Stilting, of apse, 295, 
String course, 134 f, 146, 161, 236. 
Strip-work, 98, 273, 277. 
Structura caementicia, 3. 
Stubbs, bishop, quoted, 45 ; referred 

to, 219, 333. 
Suevres, 50. 

Sundials, Saxon, 77, 115, 131. 
Swithun, St., 34, 303, 333. 

T shaped plans, 241, 283, 286. 

Table-altar, 12. 

Tamworth, herring-bone work at, 71.* 

Taper, in German shafts, 65 ; in 
Roman shafts, 9, 258. 

Tasburgh, Norfolk, 161. 

Temples, Roman, 8 ; supposed, at St. 
Pancras, 125. 

Tesserae, 10, 12. 

Tewkesbury, Norman balusters at, 198. 

Thickening of walls for a tower, 217, 
219, 221, 224, 238. 

Thinness of walls, 85, 296. 

Thorn, W., quoted, 125. 

Through-stone, 63, 91*. 

Throne, episcopal, 262. 

Tickencote, Rutland, 171. 

Timber work, 37 ff. 

Tintagel, Cornwall, 80. 

Tombstones, carved, 77. 

Tooling, 4, 5, 6,* 7, 9. 

Tours, 46, 285, 323-4. 

Towers, 82, 157, 285 ; axial, 217 f., 
286 ; central, 238 f, 240,* 242, 
286, 306 f ; defensive, 56, 304 f; 
detached, 285 ; Irish round, 36, 
166, 183, 305 ; Lincolnshire, 
158 ff., 305; Norman, 56, 159; 
pavilion, 285; round, 182 f.; 
twin, 55-6, 243*; western, 54, 
56, 82, i57f, 242, 275, 285 f, 
289, 299, 304. 



Transitional forms, 298, 300, 326. 

Trapezoidal imposts, 107-8,* 181, 326. 

Transept, 14, 16, 18, 232, 235 f., 240, 
293, 295. 

Triangular-headed openings, 27, 31, 
224, 302. 

Trier, details from, 8, 60, 62,* 64,* 
67,* 202.* 

Trinite, Stc. Caen, 179, 269. 

Tudor, C. L. R., referred to on Kirk- 
dale, 106. 

Tugby, Leicestershire, 63. 

Tumuli, Irish, 19. 

Turret, semi-circular stair, 175, 211, 
221, 247. 

Twisden, referred to, 317, 319. 

Uriconium, 8, 195, 255. 

Van de Putte on Blandinium, referred 

to, 243. 
Vaults, 7, 127, 143, 169, 266, 270. 
Ver, Normandy, tower at, 159, 162. 
Verona, 65. 

Vieux-Pont-en-Auge, Normandy, 49. 
Vikings, see * Danish inroads.' 
Villas, Roman, 296. 
Vitale, San, Ravenna, 22, 321 f. 
Volute caps, I76f., 180.* 
Vorhalle at Lorsch, 50, 59,* 67, 301. 
Voussoir-shaped bricks, 228. 

Wales, early architecture in, 18, 80. 

Warelwast, bishop of Exeter, 313. 

Weaverthorpe, Yorks, 160. 

Wendens Ambo, Essex, 63. 

Werden, a.d. Ruhr, 34, 53, 61, 66-7,* 
212, 263, 276, 306. 

Western ; apses, 55, 260*-!, 263 ; 
choirs, 53f. ; divisions in naves, 
217, 219 f., 224; doors to towers, 
164, 166 ; forebuildings, in Ger- 
many, 53 f, 304 ; towers, see 
' Towers, western.' 

Westminster, 50, 68, 81, 313. 

Westphalia, 54-5, 276. 

Westwell, Kent, 128, 178. 

Westwerke, German, 53, 304. 

Whetstone, stones of churches used as, 

White, Hugh, referred to, 247, 274. 

Whiterne, 294. 

Wilfrid, his architectural achievement, 
3 1 7 ff. ; see also ' Hexham.' 

William of Malmesbury, 79, 129, 313; 
Gesta Pontificum^ quoted, 132, 
313, 314, referred to, 34, 322 ; 
Gesta Regum, referred to, 34, 314. 

Willebrord, 46. 

Willis, Professor, quoted, 44 ; referred 
to, 157, 244-5, 257» 260, 267, 

Wiltshire Magazine, quoted, 133, 139, 
referred to, 1 74. 

Winchester ; cathedral at, Norman, 
269, Saxon, 245, 257, 267, 313, 
320 ; volume of Archaeological 
Institute referred to, 245, 257, 
267, 320. 

Windows, 90, 93*-4* ; double, 44, 60 
f , 62,*9i*-2,*236, 273; double- 
splayed, 23, 65 f, 74, 82, 92-3*, 
138, 275 ; Early Christian, 65 ; 
German, 60 f.; internally splayed, 
31, 66, 115,* 146, 273 ; Irish, 
23, 27, 31 ; keyhole, i65*-6 ; 
Norman, 82, 273 ; Roman, 8 ; 
Saxon, passim ; splayed, 8, 66, 
122 ; square-headed, 27, 115.* 

Worcester, 269, 312 f 

Wroxeter, see ' Uriconium.' 

Wulfstan, bishop of Worcester, 3 1 2 f. 

Wiirzburg, 284. 

York, 323 ; Saxon cathedral at, 34, 245, 
257, 319^- 

Zeno, San, at Verona, 65, 271. 




II III iiii I 

AA 000 628 290 9