Brooches in Late Iron Age and Roman Britain

By D. F. Mackreth

The result of forty years of study, this book offers an overview of the most common find, after coins, on sites in Roman Britain, the brooch. Used basically to hold outer clothing together, it was always on view and was usually decorative. Based on the study of some 15,000 specimens, the second volume illustrates some 2,000, all drawn by the author. The first chapter is a discussion of manufacturing techniques, methods of study and the concept of dating. The bulk of the book consists of nine chapters examining in detail the myriad style of brooches from the second century B.C., when the habit of wearing brooches really took off, to the early fifth century A.D. when newcomers brought their own types of brooch and imposed them on the rest of what was to become England. The final chapter is a synthesis of various strands mentioned in the body of the book and the social implications of the great change in brooch wearing which occurred in the third century.

• Language: English
• Publisher: Oxbow Books
• Release date: Aug 8, 2011
• ISBN: 9781842176429

Book preview

Volume 1

Contents

Chapter 1. Introduction

Part 1. The Study

Part 2. Dating

Part 3. Typologies and Classification

Part 4. Selection and Bias

Part 5. Materials and Manufacture

Part 6. The Illustrations

Chapter 2. Late La Tène, Britain and the Continent

Part 1. The Stead, Birdlip, Nauheim and Drahtfibel Group, etc.

Part 2. The Rosette and Langton Down Group

Part 3. The Colchester

Part 4. The Aesica

Part 5. The South Western La Tène Series

Part 6. The Military La Tène II

Chapter 3. The Colchester Derivative

Part 1. The Harlow Spring System

Part 2. The Rearhook Spring System

Part 3. The Polden Hill Spring System

Part 4. The Hinged Pin

Part 5. Polden Hill/Hinged Pin

Chapter 4. The Headstud and others

Part 1. Alternative Headstuds

Part 2. The Headstud

Part 3. The Wroxeter

Part 4. Colchester Derivatives, with Trumpet-style Knops

Chapter 5. The Trumpet and its Varieties

Part 1. Mainstream Trumpets

Part 2. Double-lugged

Part 3. The Knop Replaced by Flat Plates

Part 4. Hinged

Chapter 6. Continent Imports and Their Influence

Part 1. Alésia-Aucissa Series

Part 2. The Hod Hill

Part 3. The Durotrigan Type

Part 4. The Augenfibel and Relatives

Part 5. The Pannonian, Norican etc.

Chapter 7. The Plate and Related, and Dragonesques

Part 1. British

Part 2. Continental

Part 3. Objects and Animals

Part 4. Dragonesque

Chapter 8. The Knee, Almgren 101 and Interlopers

Part 1. The Knee

1.a. British

1.b. Continental

Part 2. Almgren 101

Part 3. Interlopers from Free Germany etc.

Chapter 9. The Crossbow Sequence

Part 1. The Sprung-pin or Proto Crossbow Brooches

Part 2. The Crossbow and its Antecedents

Chapter 10. Penannulars

Part 1. Coiled

Part 2. Folded Over

Part 3. Knobbed

Part 4. Late-zoömorphic

Part 5. Others

Chapter 11. Usage, Tribes, Fashions and the Demise of the Bow Brooch

Part 1. Who Wore Brooches, Why and How

Part 2. The Problem of Military Brooches

Part 3. Religion

Part 4. Marketing and Money

Appendices

1. The Dating of the King Harry Lane Cemetery

2. The Dating of Applied White Metal Trim

3. South Cadbury South West Gate

Bibliography

Volume 2

Contents

Figures

Figure 1 A Map of England showing the abbreviations used for counties in the database

Figure 2 The stamps used on the Type 3 gilded oval/round brooches

The Plates

Plate 1. Chapter 2, Part 1, 1.a1–1.b1

Plate 2. Chapter 2, Part 1, 1.b1–1.b3, 2.a–2.b

Plate 3. Chapter 2, Part 1, 2.b, 3.a–3.c

Plate 4. Chapter 2, Part 1, 3.c–3.x2, 4.1a

Plate 5. Chapter 2. Part 1, 4.1b–4.2a

Plate 6. Chapter 2, Part 1, 4.2c, 4.3, Nauheim, 1a–1.b

Plate 7. Chapter 2, Part 1, ND, 1.a–2.d

Plate 8. Chapter 2, Part 1, ND 3.a1–3.b3

Plate 9. Chapter 2, Part 1, ND 3.b3–4.b2

Plate 10. Chapter 2, Part 1, ND 4.c–Odd 3.1a

Plate 11. Chapter 2, Part 1, Odd 3.1b–Odd 4, D 1.a–1.b, DD 1a

Plate 12. Chapter 2, Part 1, DD 1.b1–7C+.1

Plate 13. Chapter 2, Part 1, DD 7C+.2–Almgren 16

Plate 14. Chapter 2, Part 1, Almgren 16–IA Odd 3

Plate 15. Chapter 2, Part 1, IA Odd 4–Odd x, Part 2, ROS 1.a–1.b

Plate 16. Chapter 2, Part 2, ROS 1.b–4.a

Plate 17. Chapter 2, Part 2, ROS 4.a–LEO 5.d

Plate 18. Chapter 2, Part 2, LEO 5.d–8 Odd

Plate 19. Chapter 2, Part 2, 8 Odd–LD 1.b

Plate 20. Chapter 2, Part 2, LD 2.a–4

Plate 21. Chapter 2, Part 2, lD 5–10, Part 3, C 1

Plate 22. Chapter 2, Part 3, C 1–C 3.1x

Plate 23. Chapter 2, Part 3, C 3.1x–C 4.b

Plate 24. Chapter 2, Part 3, C 4.b–C 4.d

Plate 25. Chapter 2, Part 3, C 4.d–C 6.b

Plate 26. Chapter 2, Part 3, C 6.c–C 7.ab

Plate 27. Chapter 2, Part 3, C 7.ab–C Odd 2, Part 4, AES 1

Plate 28. Chapter 2, Part 4, AES 1–AES 2.c

Plate 29. Chapter 2, Part 4, AES 2.c–AES 4.a

Plate 30. Chapter 2, Part 4, AES 4.b–AES 4.x, Part 5, SW, Part 6, MIL La T II

Plate 31. Chapter 2, IA Dreg. Chapter 3, Part 1, CD Ha 1.a1–1.a3a

Plate 32. Chapter 3, Part 1, CD Ha–1.a3b–1.c

Plate 33. Chapter 3, Part 1, CD Ha 1.d–2.d

Plate 34. Chapter 3, Part 1, CD Ha 2.d–3.b

Plate 35. Chapter 3, Part 1, CD Ha 3.c–5.a

Plate 36. Chapter 3, Part 1, CD Ha 5.b–X.a

Plate 37. Chapter 3, Part 1, CD Ha X.b–H.2

Plate 38. Chapter 3, Part 2, CD RH 1.a–1.d

Plate 39. Chapter 3, Part 2, CD RH 1.e–2.h

Plate 40. Chapter 3, Part 2, CD RH 2.i–4.c

Plate 41. Chapter 3, Part 2 CD RH 4.d–7.a

Plate 42. Chapter 3, Part 2, CD RH 7.b–Odd

Plate 43. Chapter 3, Part 2, CD RH Odd–4/3

Plate 44. Chapter 3, Part 3, CD PH 1.a–1.d

Plate 45. Chapter 3, Part 3, CD PH 1.X–2.b2

Plate 46. Chapter 3, Part 3, CD PH 3–4.c

Plate 47. Chapter 3, Part 3, CD PH 4.c–4.i1

Plate 48. Chapter 3, Part 3, CD PH 4.i2–4 oddments

Plate 49. Chapter 3, Part 3, CD PH 5.a–5.a3

Plate 50. Chapter 3, Part 3, CD PH 5.a3–5.c

Plate 51. Chapter 3, Part 3, CD PH 5.d1–6.a6

Plate 52. Chapter 3, Part 3, CD PH 6.a7–6.b6i

Plate 53. Chapter 3, Part 3, CD PH 6.b6i–9.c

Plate 54. Chapter 3, Part 3, CD PH 9.C–CD dregs, Part 4, 1.a–1.b

Plate 55. Chapter 3, Part 4, CD H 1.c–3.a, 2545 = CD PH 6.a3

Plate 56. Chapter 3, Part 4, CD H 3.b–4.a

Plate 57. Chapter 3, Part 4, CD H 4.b–4.k

Plate 58. Chapter 3, Part 4, CD H 4.k–6

Plate 59. Chapter 3, Part 4, CD H 6–8a

Plate 60. Chapter 3, Part 4, CD H 8b–10.a

Plate 61. Chapter 3, Part 4, CD H 10.b–10.d

Plate 62. Chapter 3, Part 4, CD H 10.e–Oddments

Plate 63. Chapter 3, Part 4, CD H Oddments–Oddments X

Plate 64. Chapter 3, Part 4, CD H 13.a–15.b

Plate 65. Chapter 3, Part 4, CD H 15.c–17.a

Plate 66. Chapter 3, Part 4, CD H 17.a–X.2

Plate 67. Chapter 3, Part 4, CD H X.2–CP, Part 5, 1.a–1.b1

Plate 68. Chapter 3, Part 5, CD H 1.b1–Odd X2

Plate 69. Chapter 3, Part 5, CD H Odd X2.

Chapter 4, Part 1, Proto HDST 1.a–1.c

Plate 70. Chapter 4, Part 1, Proto HDST 1.c–1.ga

Plate 71. Chapter 4, Part 1, Proto HDST 1.ga–1.x, Part 2, Headstud 1.a–2.c

Plate 72. Chapter 4, Part 2, Headstud 2.c–5.a

Plate 73. Chapter 4, Part 2, Headstud 5.a–10a

Plate 74. Chapter 4, Part 2, Headstud 10.a–R1

Plate 75. Chapter 4, Part 2, Headstud R1–X2

Plate 76. Chapter 4, Part 3, WROX 1.a–1.b

Plate 77. Chapter 4, Part 3, WROX 1.b–1.c, Part 4, CD H/PH 1 a–e

Plate 78. Chapter 4, Part 4, CD H/PH 1 x–xx.

Chapter 5, Part 1, TR 1.a1a–1.a1b

Plate 79. Chapter 5, Part 1, TR 1.a1c–1.2a

Plate 80. Chapter 5, Part 1, TR 1.2b1–1.2bx

Plate 81. Chapter 5, Part 1, TR 1.2by–1.2cx

Plate 82. Chapter 5, Part 1, TR 1.3a1–1.4

Plate 83. Chapter 5, Part 1, TR 1.5a–2.1

Plate 84. Chapter 5, Part 1, TR 2.2a–2.2f

Plate 85. Chapter 5, Part 1, TR 2.2g–3.1b

Plate 86. Chapter 5, Part 1, TR 3.1b–4.1b2

Plate 87. Chapter 5, Part 1, TR 4.2–x2.b

Plate 88. Chapter 5, Part 1, TR x2.c–TR Dregs

Chapter 6, Part 1, Alesia-Auc 1.a–1.b1

Plate 89. Chapter 6, Part 1, 1.b1–2 AVCISSA 2x

Plate 90. Chapter 6, Part 1, 2 AVCISSA 2.a–3.b, Part 2, HOD HILL 1.a

Plate 91. Chapter 6, Part 2, HOD HILL 1.a1b–2.b

Plate 92. Chapter 6, Part 2, HOD HILL 2.b–4.a2

Plate 93. Chapter 6, Part 2, HOD HILL 4.a2–4.c1

Plate 94. Chapter 6, Part 2, HOD HILL 4.c2–4.d3

Plate 95. Chapter 6, Part 2, HOD HILL 5.a–7

Plate 96. Chapter 6, Part 2, HOD HILL 7–10.b

Plate 97. Chapter 6, Part 2, HOD HILL 10.c–12.a3

Plate 98. Chapter 6, Part 2, HOD HILL 12.b–13.a2

Plate 99. Chapter 6, Part 2, HOD HILL 13.a2–13.b, Part 3, DURO 1.a–1.b

Plate 100. Chapter 6, Part 3, DURO 1.b–4

Plate 101. Chapter 6, Part 3, DURO 5–6.c

Plate 102. Chapter 6, Part 3, DURO 6.d–8.b, Part 4, AUGEN

Plate 103. Chapter 6, Part 4, AUGEN 1.1–Oddments, Part 5 PAN 1.1–1.2

Plate 104. Chapter 6, Part 5, PAN 1.3–1.6, Norican

Chapter 7, Part 1, British, Plate 1.a–1.f

Plate 105. Chapter 7, Part 1, British, Plate 1.a–1.f

Plate 106. Chapter 7, Part 1, British, Plate 1.f–3.a3

Plate 107. Chapter 7, Part 1, British, Plate 3.a4–3.b6 ii

Plate 108. Chapter 7, Part 1, British, Plate 3.b6 iii–6.d

Plate 109. Chapter 7, Part 1, British, Plate 6.3–7

Plate 110. Chapter 7, Part 1, British, Plate 8.a–unclassified, Part 2, PL CONT 1.a1–1.a5

Plate 111. Chapter 7, Part 2, PL CONT 1.b–2.a

Plate 112. Chapter 7, Part 2, PL CONT 2.a–2.d1

Plate 113. Chapter 7, Part 2, PL CONT 2.d1–3.x

Plate 114. Chapter 7, Part 2, PL CONT 3.x–4.x

Plate 115. Chapter 7, Part 2, PL CONT 5.a–7.c

Plate 116. Chapter 7, Part 2, PL CONT 8.a–12

Plate 117. Chapter 7, Part 2, PL CONT 13.a–17.a

Plate 118. Chapter 7, Part 2, PL CONT 17.b–20.3b

Plate 119. Chapter 7, Part 2, PL CONT 20.3c–20.6x

Plate 120. Chapter 7, Part 2, PL CONT 20.6x–22

Plate 121. Chapter 7, Part 2, PL CONT 23.a–UNC 1

Plate 122. Chapter 7, Part 2, PL CONT UNC 1–UNC 3

Plate 123. Chapter 7, Part 2, PL CONT UNC 3, Part 3, OBJECT 1.1a–1.2b1

Plate 124. Chapter 7, Part 3, OBJECT 1.2b2–2.2

Plate 125. Chapter 7, Part 3, OBJECT 3.a1–3.d

Plate 126. Chapter 7, Part 3, OBJECT 3.f–4.d2

Plate 127. Chapter 7, Part 3, OBJECT 4.dx–7.1

Plate 128. Chapter 7, Part 3, OBJECT 7.2–7.4, Part 4, DRAG 1.a–1.b

Plate 129. Chapter 7, Part 4, DRAG 2–3.b

Plate 130. Chapter 7, Part 4, DRAG 3.b–3.x1.

Chapter 8, Part1, KNEE Br 1.a1–2.a1

Plate 131. Chapter 8, Part1, KNEE Br 2.a2–5, KNEE Cont 1.a–2.a

Plate 132. Chapter 8, Part1, KNEE Cont 2.b–XX

Plate 133. Chapter 8, Part 2 Almgren 101 a–e, Part 3 FG FG.a–FG.c

Plate 134. Chapter 8 Part 3 FG FG.d–FG.f.

Chapter 9, Part 1, proto CR 1a–2a

Plate 135. Chapter 9, Part 1, proto CR 2.a–3.d

Plate 136. Chapter 9, Part 1, proto CR 3.x, Part 2 CR 2.a–2.d

Plate 137. Chapter 9, Part 2 CR 2.e–3.2b

Plate 138. Chapter 9, Part 2 CR 3.2b

Plate 139. Chapter 9, Part 2 CR 3.2b–3.4a1

Plate 140. Chapter 9, Part 2 CR 3.4a1–3.4b

Plate 141. Chapter 9, Part 2 CR 3.4c–3.5

Plate 142. Chapter 9, Part 2 CR 3.5–Oddments 2

Plate 143. Chapter 10, Part 1, PEN c1.a–c2.c

Plate 144. Chapter 10, Part 1, PEN c2.d1–c2.d3, Part 2, PEN f1.a–f.2a

Plate 145. Chapter 10, Part 2, PEN f.2a–f.f, Part 3 PEN k1.a–k1.b

Plate 146. Chapter 10, Part 3, k1.b–k3.b

Plate 147. Chapter 10, Part 3, k4–k Oddments, Part 4, 6

Plate 148. Chapter 10, Part 4, 6–7

Plate 149. Chapter 10, Part 4, 7–9

Plate 150. Chapter 10, Part 4, 9–XX

Supplementary Plate 1

Chapter 7, Part 1, Plate 1 the rest, 3.a, 3.a1, 4, Part 2 PL CONT 2.c2, 4.b, 8.a, 12, 20.x, UNC 3

Supplementary Plate 2

Chapter 7 Part 3, 3e

Chapter 1. Introduction

Part 1. The Study

All the index cards were arranged initially by the way in which the pin was mounted. As I had only intended to study bow brooches, the Colchester came first, then followed the chief methods used in the Derivative forms and these were arranged by apparent increasing complexity. Therefore, the Rearhook came before the Harlow and that preceded the Polden Hill. Being a firm believer in typology and a sort of degeneration, I placed the hinged-pin brooches last. We will see how foolish this was in terms of absolute typology and how the degenerative principle can safely be ignored. Within each major group, I arranged the brooches in order of increasing complexity until, like St. Paul, the scales fell from my eyes. I have had few such moments of illumination, and only in earthly terms.

One such was when I discovered that Clarke's discussion of polythetic assemblages ([146] [] = no. in Bibliography, 37–38, 293–306) not only had relevance to brooches, but was actively useful. Needless to say, what is true of one group, type, sort or variety is not necessarily true of another. Certain fundamentals only became obvious when there were enough examples and I had begun to look in greater detail at the major publications of continental material. This is best exemplified in Plate brooches in which bilateral springs mounted either on a single or between double lugs (see Chapter 7) proved to be British by their much greater frequency here than on the continent where most with springs can be shown to be British types. This carries with it the entirely British phenomenon of white metal beading or rosettes (see Appendix 2). These were applied to the surface of an enamelled brooch so that the colours glint jewel-like in the depths of the silvered finish.

I only met Mr. Hull once, one weekday morning early in 1966, at the British Museum and he volunteered the information that he only really undertook the study of brooches in the hope that different types might define the areas occupied by individual British tribes based on what were then termed cantonal capitals. He said that he soon found that was not going to happen, but by then he had embarked on a major study and was not inclined to give it up.

My intentions were less high-minded, being concerned with the dating of the most common class of object found on sites after coins. The material from the Sheepen published by Hawkes and Hull [385] could be said to have had a limited time-span, but there was little guidance as to how accurate it was: what was common before the conquest and after c. 65 was far from clear and, as subsequent work has shown, could not be adequately reflected on that site alone. Similarly, Collingwood's approach to dating [158] was heavily influenced by his philosophical viewpoint that brooches of good design are followed by cheap imitations. Ones views are inevitably coloured by a subjective approach when faced with something outstanding like the Aesica Brooch. Definitely of first-class quality by whatever criteria one may choose, there can be different levels of craftsmanship within the mass of Aesica-Type brooches. Quality of finish could have been related to price and some products of a genuinely untalented worker should be expected in the archaeological record.

However that might be, I was primarily interested in establishing proper date-ranges while the size of the index remained small and in building up a framework for a basic typological study. It was, however, only with the increase in the publcation of pre-conquest material that the relationship of truly Iron Age products with well known post-conquest brooch types could be seen. It was by looking mainly at the technical aspects of brooches that I could see that the very things that Hull had hoped to find was possible, not so much individual tribes, but culturally discrete assemblages dating back at least to the beginning of the first century. These are discussed in Chapter 11, Part 1.

In justifying the division of brooches on their technical aspects, one detail perfectly illustrates that no matter what influences worked on the artisan in producing a saleable article, the basic way in which the pin was fixed was so deeply ingrained in him that, as it seldom has any visual value, it would by preference be the method he was first taught. A demonstration of this lies in brooches with bilaterally sprung pins. When the brooch is viewed from the front, the spring coils out to the left and then returns via the chord to the extreme right-hand end to coil back to the centre and turns down to end in the pin. The alternative direction is taken by less that 1% of the brooches with this type of spring. Therefore, an absent spring can safely be restored this way. Even if nearly all Rearhooks (Chapter 3, Part 2) have lost their springs, it is noticeable that solder, or traces of it, is found behind the left-hand wing and occasionally behind the right-hand one. In other words, the spring, as far as the craftsman was concerned, had to return via the chord from left to right. This detail is to be seen on integral springs and on all Trumpets with a single lug system. If a brooch spring lies in pieces, the restorer should always follow the rule that the spring coils out to the left first and the fragments will fall into place logically, providing not too many parts are missing. Whether the chord was external or internal the fragments often reveal this as well.

Part 2. Dating

This diagram seeks to show two principal things: a is when a brooch is first made, b is when it passes out of manufacture; c represents the point at which the brooch in manufacture begins to enter the archaeological record in sufficient numbers for an initial date to be arrived at by ordinary analysis of site finds; d represents the time after which only the rarest of survivors-in-use will still be seen. It follows that the general archaeological dating for a particular brooch is c–d. Thus, the true date for points a and d will be hard to assess.

Most brooches are found as part of the cultural debris on ordinary habitation sites and so must be treated as any other class of material. However, one class of site which can produce brooches has to be treated differently: cemeteries. The objects deposited as either grave goods or on the body follow different rules. A brooch deposited on the body itself may well bear a different relationship to the grave than one merely placed by the body. This, of course, should only really apply to inhumation burials, on the grounds that almost all material found in furnished cremations was clearly not in the pyre. In general, it should be accepted that grave goods stand more chance of being a purely contemporary group than miscellaneous site rubbish which will include differing quantities of residual material. Arguments based on the proposition that grandma's out-moded brooches would be discarded first should really only apply to graves containing grandmas: why should her belongings be held over to be placed with someone else and from whence would her grave goods have come? Although the dating used to arrive at the time when the burial was made, will be based upon normal archaeological criteria, the reality is that graves stand a better chance of falling with the period represented by a–b than c–d. But it must be admitted that there should be a fair number of graves which will actually belong to the period b–d, but they should be marked by a higher proportion of old and worn gear than others in the cemetery.

Furnished Iron Age and Roman cemeteries may, on the whole, be uncommon in the archaeological record, especially those which are large enough for some kind of internal chronology to appear. There are only two which really count at the moment: the King Harry Lane cemetery [675] and Lankhills, Winchester [147]. I shall have cause to discuss the dating of the former (Appendix 1) and the make-up of brooches in the latter (Chapter 11, Part 2).

The text is peppered with numbers in () brackets; these are to specific brooches in the chief database each of which has a unique number. Sundry bar charts appear and these were constructed by assessing the length of time during which a brooch could have been deposited, in short the overall date of its context. The series is arranged in order of the earliest date of the context, the whole period appearing as a column with that base. Hence brooches from the Roman fort at Newstead are represented by bars running from 80–200, one for each brooch, the 80 determining where it stands in the whole of the group under examination.

In other words, the earliest date on the diagram stands a chance of representing point a in the diagram above. While that could lead to some discussion, what the best of the bar charts shows is the point at which brooches cease to be a regular part of the archaeological record. An orderly progression usually breaks down, frequently in the late second century or the early third and thereafter there are sporadic occurrences. Arguments based on the disruption of the archaeological record due to the difficulty of dating in the third-century have a certain validity, but it is noticeable how seldom a regular image emerges when dating becomes more reliable. It is the use of these charts which has enabled me to be fairly definite about the overall floruit of particular brooch types, and which demonstrates the peculiar nature of the terminal date of the Rearhook.

In the serial list given on the CD-ROM, the publication source of that particular brooch is given in the publication column a number in [ ] brackets which is that given to the relevant entry in the Bibliography

However, an element must remain about the Iron Age end of the spectrum. It is true that far more Roman sites have been excavated than Iron Age ones, although the gap is being narrowed. Iron Age material is bedevilled by a general lack of finds, compared with an equivalent Roman site away from the highland zones of Britain, compounded by what was a more effective system for recycling metal than was applied to the first 150–200 years of the Roman occupation.

Part 3. Typologies and Classification

The bug-bear for new students is the almost bewildering array of attributes which brooches can have. A purely formal approach (e.g., [617]) will divide the available corpus into major groups, and number these, then subdivide into lesser major groups, and number those, and then finally detect sub-varieties which will also be numbered. The result is an intellectual triumph, but not memorable: how many people actually manage to remember more than, say, ten Almgren types? The difficulty in using standard terms is that many of these have a wide currency, but do not necessary provide a clear indication of their origin. The Hod Hill is one and the Langton Down ([756], 71–4, fig. 10) is a notorious other. Both are continental in manufacture, but neither has a satisfactory other name. As this book is about brooches in Britain, I shall unashamedly retain both names. Thereafter, we have Colchesters and a few foreign names which are useful, such as Augenfibel, Nauheim and Drahtfibel (Wire Brooch), and these are also kept.

Of the three major divisions of brooches found in Britain, the earliest is Collingwood's alphabetical system which is rightly castigated, the next would have been a numbered series devised by the late Rex Hull and the last is one devised by Richard Hattatt, an indefatigable worker blessed with great perception. He tended to use names and accepted numbers, hence Birdlip and Type G Penannulars, but there is a limit to how memorable such a scheme would be if applied to a greater body of information, or even how the names should be arrived at. The student is presented with a dilemma and my way out of this morass is to use a combination of names and numbers, which, I fear, will prove to be intractable to many.

I have opted for names for major families, Colchester Derivative, Nauheim Derivative and the like, and names also for major sub-groups which cover scores of examples: here Rearhook and Harlow in Colchester Derivatives can be cited. Thereafter, numbers come into play as ultimately there is a residue in each major family difficult to divide into groups of truly associated versions in a single polythetic family. Future work by others should be able to refine by adding more examples and, with a bit of luck, the residue will become smaller and smaller as a percentage of the available corpus, even if actual numbers remain obstinately the same. The whole of the system is not set out here but will be found in Chapters 2–10 where the types and varieties are discussed.

Lastly, there is inevitably a residue of brooches which refuse to fit any category. These have been sorted into their most likely chapter and appear at the end of each as Desperate Cases.

The reader, skimming through Chapters 2 to 10 may well be daunted by the enormous number of divisions which appear, but help is at hand in the Database. The simple names devised can in some instances be reduced to initials all of which derive easily from their designation. Thus Rearhook becomes RH, Colchester Derivative is CD and as for the Nauheim and the Drahtfibel Derivatives, these are ND and DD respectively. This aide-memoire is necessary as both have a Type 1 and it is only the letters of the abbreviated form which distinguishes them on the accompanying CD-ROM. As the basic database has a sort engine, it will be easy to isolate all the examples of any particular major Type and each sub-variety. The same engine can be used for arriving at site assemblages. The CD-ROM has a simple listing of these abbreviations and the chapters and sub-sections where the relevant items are to be found.

Part 4. Selection and Bias

There has been little exclusion from what was to go into the Corpus beyond the obvious one of fragments of spring and pin. The major bias in selection caused by not dealing with Plate brooches and Penannulars from the beginning became statistically less significant as the index grew.

However, I have had no real choice over the areas which have favoured me with material for report. Taking East Anglia, I have seen a lot of brooches recovered by metal detectors and which came to me through the good offices of Bill Milligan of the Castle Museum Norwich. I have picked up, inevitably, a fair amount of major published material from Essex, but Suffolk in comparison is like a black hole. Otherwise, the rest of England south of the Dee–Humber line is covered patchily but probably effectively, which is more than can be said for the lands north of that line. I have material from parts of South Wales, but even allowing that Wales has produced little, North Wales is under represented. Margaret Snape has kindly made available the items recorded by her from the forts of the Stanegate and there are groups of useful material from parts of Hadrian's Wall. However, there is a major imbalance in the assemblages north of the Dee–Humber line: nearly all of it comes from Roman military sites, the Caves of Derbyshire and Yorkshire being honourable exceptions.

Whereas in the southern parts of England there is a fair mixture of material from town, villa and purely rural settlement, assemblages from the north come firstly from forts and a lagging second from town or vicus. Practically nothing has been recorded from plain rural sites. It is possible that this really represents the quantity of material to be expected from such sites, but the truth is more likely that the information is masked by a lack of excavation. This may be a serious omission and as such severely limits the value of any discussion dealing with brooch wearing in the third and fourth centuries, if not earlier.

Practically all items in the Corpus have been published in one way or another – individual excavation reports, papers in journals, or as notes – and this is a surprising state of affairs given that so little had been published before 1960. In other words, there has been a revolution in how small finds are treated. It is no longer acceptable to publish excavation reports, such as those on the Great Casterton town and villa, in which a few items of interest to the editor get a mention and the bulk is passed over in silence. Even if financial constraints prevent full publication in letterpress or microfiche, there is usually a statement saying where an archive can be consulted. But, and it is a big one, outside fortunate publications, what is selected for illustration can be quite arbitrary. If the item is enamelled, it is often described in conversation as being pretty or nice, so much so that the latter term has become, for me, shorthand for a Plate brooch of second-century date. Similarly, the Nauheim and Drahtfibel Derivatives tend to get short shrift. The matter is aggravated when it comes to individual notes, which tend to be of the An unusual triple-bowed exotic zoömorphic fibula type. If one relies only on publications, then this sort of bias is bound to affect the perception of what is common. It is assemblages of common brooches which need to be published. In fact, one report of mine [495] was nearly turned down for a publication grant on those grounds, yet it is precisely because they were ordinary that they needed to be published in a region in which publications had failed to show what was commonplace.

The same general point can be made about museums where, outside collections derived from modern excavations, the dead hand of taste had removed virtually all iron brooches and the Nauheim and Drahtfibel Derivatives as they were visually unattractive and hence not worth exhibiting. The British Museum Collection, outside its site-specific assemblages, is a glaring example of one built up on the basis of attraction rather than being typical. As can be imagined, it has a much higher percentage of enamelled brooches than any site assemblage would warrant. Many another type will have suffered by being passed over on no rational basis, except what was attractive to a museum curator who may well have been more familiar with medieval alabasters, nineteenth-century corsets or every variety of gun from one manufacturer than with any archaeological material. Those of us who have been known to get our hands dirty are probably still regarded by museum curators as scrabblers in mud: an otherwise admirable man in charge of the Castle Museum at Nottingham said in a public address Archaeologists concern themselves with things which our ancestors, quite rightly, threw away.

My greatest regret is that there is now almost no chance of building up a corpus which includes the bulk of brooches above ground. In the old days, this was a feasible proposition, given a private income, as nearly every brooch lurked in a museum collection and, if from an excavation, was likely to be published. However, the explosion in the use of metal-detectors, despite the protestations of the highest of motives, has meant that sites are daily plundered and the number of unprovenanced brooches is going to become a severe embarrassment to the student in the future: I have details of some private collections, and the sites which have produced the items which, in fifty years' time, will probably be the only record. Many museums see items collected in this way and there are people who record in one way, frequently badly, what is brought in, but there is no time to travel round every museum and pick up these details. The Portable Antiquities Scheme is a brave attempt.

Part 5. Materials and Manufacture

In theory any material which can be made stiff enough to form a bow and pin can be used but, in reality, only metals are known, even lead (Chapter 3, Part 3). The commonest surviving metal is a copper alloy, the next commonest is iron and the rarest are, naturally, gold and silver. Iron is almost certainly under represented in the archaeological record, its manner of corrosion easily reducing it to shapeless lumps of basic soil bound by migrated iron compounds. X-ray is almost always needed to establish what an object had been and it is only in recent times that collections of corroded iron have been subjected to systematic study with the consequent increase in the numbers of known iron brooches. However, it can be demonstrated that the extensive use of iron was a cultural phenomenon quickly coming to an end in the Early Roman period, possibly as the result of the change from forging brooches in favour of casting. It would have been harder to produce a sophisticated shape this way, yet the Involute series of La Tène II Brooches ([415], 156–70, pls 46–9) shows that there were no fears or clumsiness in doing just that. The finished product would have gleamed like silver at a distance and may have had to have been oiled to prevent rust.

Copper alloys have been subjected to the greatest study, and there is no need to discuss them as they have been more than adequately covered by Justine Bayley ([45], 12–25) who has also dealt with the basic decorative techniques used (ibid. 26–51). It is worth adding, however, that enamelling is known on iron, both in the Iron Age in the case of swords ([673], 66–70, figs 51–3, frontispiece) and in one workshop in the Roman period ([71], 26, 31–4, 84).

The chief types of brooch forged from rolled or folded sheet are, naturally, the simplest forms: the Penannular and the basic Nauheim or Drahtfibel Derivatives. It is more than likely that the technique was used for the parent types themselves, but they occur so seldom that the chance to check this in Britain has not really arisen. The technique, however, was used for a major British school operating in the South West (Chapter 2, Part 5). Certain other types (Chapter 3, Part 3, Type 2.a) would seem to be suitable for the use of the same technique, but although there is evidence for hand-finishing, they were undoubtedly cast and the cast parts of the scabbard of the Peterborough Sword (Orton Meadows: Stead in Mackreth forthcoming) show that relatively large castings with very thin sections were no problem to Iron Age craftsmen and, therefore, should not have been to those trained in the same basic tradition.

Only rarely does one come across a brooch type which can use sheet metal to form hollow sections to form the whole of the major part of a brooch (Chapter 4, Part 2, Type 1.e; Chapter 9, Type 2.6). In the first instance it is probable that the use of folded or rolled sheet metal was so extensive that the craftsmen using it did not know how to produce the sturdy sections found on many Colchester Derivatives. It is also surely significant that it is in this group that one finds false springs in which the hinged pins normal for the region were mounted. In the second, the argument is that the use of sheet metal was to reduce the weight greatly so that the brooch could be worn with a delicate textile (Chapter 11).

In the early stages in the development of the Rosette or Langton Down we come to the transition from a purely forged technique to a cast one. In British brooches the same transition shows best in the Colchester. I have not noticed a single Colchester definitely made from rolled or folded sheet metal, but most show that there had been extensive forging to achieve the finished product. The evidence fits the examples of cast Colchester blanks from Baldock ([674], 122–3 fig. 50, 160), but also suspected from Odell (excavations, B. Dix, unpublished) and Ashton (excavations, J. Hadman, S. Upex and B. Dix, unpublished). These show that the beginnings of the catch-plate are there: incipient wings; a very crudely indicated catch-plate; thick bars which were to be worked up into hook and spring. The blank was straight with the stalks for the spring and hook pointing upwards. The shape of the bow and especially the bend at the top of the profile had to be fully forged and this accounts for the tell-tale marks probably of an anvil along the back of the bow. The catch-plate frequently has its junction with the bow marked by tools which helped to form it and all the piercings have to be punched through. On a poorly finished Colchester it can be seen that these were always punched through from the right, when the brooch is viewed from the front.

The die block makes a definite appearance, as analysis of Aucissas at Augst has shown ([617], 36) the bow and the head-plate were made separately and then joined. The development of the Rosette in the second half of the 1st century B.C. should be another example of the way in which the decoration on the bow is frequently repeated on the fantail foot. However, rather than hammering the casting into the die block, it is possible that a block was stamped on to the casting itself, the first method has the block on a bench, the second has the casting on the bench instead.

The only brooch mould which fully allows the technique for casting brooches to be seen is the unused one from Prestatyn ([71], 87–8, 96, fig. 39, 20, pl. XIX). The mould demonstrates two things, firstly how to produce a series of moulds from a single pattern, and, secondly, that the result was a casting which needed extensive hand work to achieve the final result. The pattern has a projection which marks the gate by which the metal will enter the mould. The split in the mould shows that half the pattern was invested in clay and the keying for the other half was made by pressing a stick or rod into the intended joint line repeatedly so that the depressions radiate from the pattern. A parting material such as fine sand would be placed on the joint line before the second part was formed. On removing the pattern, the two parts were joined and smeared round with wet clay to bind the two.

The dried moulds would be baked to drive out any remaining moisture. A series of moulds would be cast together and it would be the merest chance that one was not used and was then preserved to be found under modern archaeological conditions. Once the castings had cooled down, the clay would be knocked off thus accounting for the tiny pieces usually found. More often than not the actual surface of the mould comes away from the body and is lost when the casting is cleaned and fettled. This term means the removed of blips arising from metal penetrating air holes in the mould, the flash along the joint lines, and the down gate. These are all excrescences on the finished product and all provide new metal for further items. It is worth the thought that the filings resulting from fine-finishing were also carefully collected, a detail which has great point when working in precious metals.

It is legitimate to ask what material was used most often for patterns. Wood is possible, clay unlikely, but the most likely because of its ease of working may have been lead. A pattern from Donington, Salop, (9825) and another from Poole Cavern are just two examples. The first was for a brooch belonging to one of the Polden Hill Dophins (Chapter 3, Part 3, Type 4.b) where the indications were for the gate to have been at what we call the foot. The other belonged to the Wroxeter type (Chapter 4, Type 1.b3: brooch 5676).

Norfolk has produced two examples of metal moulds for brooches. The more extensive find was at Old Buckenham (Plate 36) where several mould halves were found for a type of Harlow [46]. The other is a single valve which came from Felmingham (Plate 38, [47]) and was for a type of Rearhook (Chapter 3, Part 2, introduction). It may not be a coincidence that both finds are from the lands of the Iceni, a tribe which seems to have had a high level of metalworking skills. Obviously there were sets of moulds, and each was re-usable. The Old Buckenham finds, with the other scraps, could be described as part of a metalworker's baggage lost in transit. How common such moulds may have been will never be known for they were in themselves a valuable resource and usually melted down.

One way of perpetuating designs, or creating more patterns, would have been to use castings themselves as patterns. How often this took place is difficult to say. Metal contracts as is cools so any new brooch made from an old one would be measurably smaller, and lighter, than its parent. In the discussion of specific types, the sizes are given and the ingenious reader is invited to explore the possibility.

The brooch as found is usually a green colour, the result of corrosion after long burial. In some instances the finish is very smooth and polished and this is more often found on forged brooches than any other and may be related not so much to the alloy used, but to the crystal structure after forging and, possibly, annealing. Cast brooches seldom needed heat treatment to correct any brittleness. At base, the original brooch would have gleamed like gold and occasionally the exceptionally preserved remains of this have fooled people into thinking that the brooch is either gilded or even of gold. Inspection under a × 10 magnifying glass should be enough to reveal the mistake as nothing can disguise the brilliant yellow of real gold. The commonest finish applied to a brooch to remove its basic yellow colour is tinning ([45] 41–2) to give a silvery finish and it is noticeable that gold is rarely used for brooches in the first and second centuries, whereas silver is occasionally found. However, silver parcel-gilt is known [83]. Tin was itself used differentially to produce the same effect and the commonest type to show this is the Hod Hill (see Chapter 6, Part 2), but it has only seldom come to notice because so few Hod Hills survive in a good enough state for those with partial tinning to stand a chance of being noticed.

Silver Sulphide or niello was used in two main periods. The earlier is the middle of the first century A.D. where it is found providing an inlaid contrast to the popular imitation silver provided by tin. The material was applied as a powder, heated and, on fusion, polished to continue the sheen imparted by the tinning. The material is frequently met on Hod Hills (Chapter 6, Part 2) and on some Plate brooches, especially those of Feugère's Atelier C (Chapter 7, Part 3, Type 7.3). I have not found niello in use with any other finish except in its second period of favour when it was used in conjunction with gold or gilding in the fourth century (Chapter 9, Part 2) and reflected current taste in high class table ware such as that in the Kaiseraugst Treasure.

Enamelling is by far the most obvious decorative finish. There are occasionally traces that it had been used on the fully developed Langton Downs and Rosettes of the Augustan-Tiberian period (Chapter 2, Parts 1 and 2) with enough technical know-how to produce bead-rows, but there is no evidence that it had been used on pre-conquest brooches of the British Late La Tène tradition.

The earliest finish found on British brooches, other than the plain metal of the brooch itself, is tinning, and enamelling makes a fitful start in the decade 60–70 (Chapter 4, Part 2, Type 3.a). The earliest colour in use is red, in common with pre-conquest preferences, but a second colour becomes almost de rigueur very soon and the commonest contrasting one is blue, hence the use of red and blue almost as standard on enamelled Headstuds (ibid., Type 5.a) and on some types of Plate brooch such as the British Domed series (Chapter 7, Part 1, Type 6.a). A silvery finish and enamel are very seldom found together almost certainly because of the higher temperatures needed to fire the enamel paste than for the washing of the surface of the brooch with tin.

Although it is generally safe to consider enamelling as a second-century phenomenon, it is sufficiently common in the last quarter of the first century for it to be a mistake to assign all enamelling to the later period. A further development can be seen in the use of silica-based materials. In a few instances it can be seen that the enamel was prefabricated and then applied as a kind of opus sectile. In such cases, it was laid on a bed of powdered enamel which was then fused together with the base metal and the surface enamel pieces to hold all together (e.g. brooch 5619 in Chapter 4, Part 3). Red appears to have been the commonest colour used as the glue. Glass canes were also prepared and then thin-sliced to be inserted into a contrasting matrix which helped to keep them in position. The canes were sometimes only a thin rod of a single colour, frequently black or white, and were used to provide dots of colour and on one type of Plate brooch formed a hobnail pattern on the sole of an item of foot-wear (Chapter 7, Part 3, Type 1.1b, Pl. 123).

All the brooch examples I have seen with these simple glass canes are continental in origin, but this does not apply to the most elaborate form of this technique: millefiore. Canes of this were formed on the same principle to make sea-side rock: blocks of colour were combined in a crude block in the pattern required and then the block was worked so that it became very thin and long, the pattern being perfectly reduced as though the block had been through a true extrusion process. Slices of millefiore were most often floated in a matrix which provided the glue needed as well as a contrasting background (e.g., Chapter 7, Part 2, Type 8.a). Even when millefiore alternated with plain enamel, it can be seen that the latter is still acting as the fixing agent (ibid). Where the entire field is filled with millefiore, it should be assumed that a plain enamel has been used underneath to fuse the whole together.

The introduction of glass dots was obviously secondary to the initial mass use of enamel itself. It also seems to be clear that millefiore is later still and that the conversion from mixed millefiore and enamel to all millefiore is again a chronological one. In other words, the disappearance of enamel as a finish in its own right ceases to be a puzzle if it had in fact been naturally superseded by a different effect. When fashion turned against that in its turn, enamel or slices from canes, or glass in prefabricated paste gems, passed completely out of use.

The ultimate in the use of silica-based materials was the direct imitation of precious or semi-precious stones. There was, when it comes to their use on the brooches I have seen, no real attempt to fool anyone, it being the simulation alone which mattered. Two types of fake can be seen, the intaglio and the cabochon gem. Both are frequent on only one main sequence of brooches and this happens to be British, the round/oval, enamelled/gilded group (see Chapter 7, part 1, Type 3.b2). The detailed discussion makes it clear that there is a problem about what usually filled the central cell on the enamelled series: I am not sure that enough glass intaglios have been found for these to have been the normal case. As intaglios occur on the gilded part of the series, I see them marking the transition between the two parts as there can be no doubt that enamelling ceased entirely when gilding was introduced. Brooches from this end of the line nearly always have a conical glass gem in the centre, frequently black in appearance, although the glass seems always to have been a very dark red purple, sometimes marbled and rarely entirely of any other colour.

The nearest I have come to a silver-plated brooch was a Trumpet in Rowley's House Museum, Shrewsbury, but it was so coated with sticky varnish that the details of the brooch were obscured. The closest I have seen to the common use of silver as a plating medium has been in the extensive use of white metal trim on a host of brooch types, all British (see Appendix 2 for its dating). The trim is not plating as such. It is a prefabricated decoration usually in the form of beaded strips, or rosettes, or twisted wire wound into cones (Chapter 8, Part 2, passim). The decoration was soldered to the brooch whose base metal was frequently allowed to show and thus giving the appearance of silver parcel gilt. In other instances, it was applied to the borders and reserved areas on flat plate brooches whose surface was almost entirely given over to enamel. In these cases, the polished fresh enamel would have gleamed like jewels set in an encrusted setting (Chapter 7, Part 1, Type 2).

The sheet used was very thin and it is hardly surprising that it so seldom survives. Usually its former presence is detectable by the presence of strips of a dull grey wash, the remains of the solder used to hold it in place, and it is the observation of the layout of the solder which allows many schemes to be restored from the few fragmentary traces occasionally found.

Gilding is another matter. The chief techniques would be the use of mercury gilding on an almost pure copper base, and evidence has been found for its use in Britain [810]. Perhaps significantly the technique at large seems to belong to date after the early third century. Unfortunately, little proper analysis seems to have been done on the base metals of brooches of the gilded round/oval plate brooches (Chapter 7, Type 3.b) and the Crossbow sequence. Occasionally, the bare base metal of a Crossbow can be seen to have a red colour which suggests that it was basically a copper and this could make such brooches more suitable for mercury gilding. However, it seems most unlikely that the members of the round/oval gilded school were also of an almost pure copper: they all, when well enough preserved, have a tinned back and it is possible that the fronts were also tinned in the hopes that the gold would adhere more easily.

Thereafter, there are a series of mechanical decorative effects. The commonest stamps are simple dots which occur too often for any particular type to be singled out. The next most common are the concentric circles which are also too frequent for any type to be picked out, but honourable mention must be made of a hugely popular continental strain (Chapter 7, Part 2, Type 1). A series of small stamps in triangles or short propeller forms to be filled with enamel; leaves or Christmas trees to be filled with niello can be found on other Continental forms (ibid., Type 2). However, the most varied collection are those to be found on Type 3.b of British Plate brooches gathered on figure 2.

Part 6. The Illustrations

The index numbered over 15000 entries when the list was closed for this book on June 28th, 2004. It had been obvious for at least two decades that it was unrealistic to think that the index could be published as a single, fully illustrated Corpus. The end product would have been hopelessly expensive and repetition of almost identical examples inexpressibly tedious. Even so, the number of those selected for publication stands at 2093. This I regard as the minimum number required to cover all the points which I have to make and to show sufficient differing varieties. Note: those brooches on the Plates marked by * are the most common varieties so far identified. Those in the Text marked by * are the ones actually illustrated.

Ordinarily, it would be difficult to spot which are the commonest forms of main types or their main sub-varieties without constantly referring to the text. While I hope, of course, that the text will be useful, the casual reader who is hoping to get a basic grasp of the subject will need a little help. The reader will have to bear in mind that the moment the list of entries for this book is closed, some site will produce an assemblage which will identify what is currently a lone example as being a genuine sub-group or major variety. Unfortunately, the onus will be on the reader to look out for such examples. The bulk of the brooches in the index has been published, one way or another, even if an exceptionally well-equipped library is needed to find them. I have consciously tried to select for illustration examples which have not been published and which stand little chance of being so, whether from reputable excavations or from private collections. People like Richard Hattatt are all too rare and with the general despoiling of archaeological sites by users of metal detectors, private collections are now numerous and, when broken up, their contents will add to a growing body of material without provenance and, therefore, of diminished use.

The original card indices of the brooches and sites, are deposited in the British Museum. The locations, or collections, along with the names of known excavators and the bibliographical citation are on the attached CD-ROM.

A note on using the CD-ROM

The database contained on the CD-ROM that accompanies this volume is a Microsoft Access database and readers will need Microsoft Access to be able to query it. Mac users will need to load the software provided but will only be able to conduct simple searches not full queries.

Note: All dates are A.D., unless otherwise specified. Those before and straddling the Great Divide are given with both indications.

Chapter 2. Late La Tène, Britain and the Continent

Plates 1–30

Here are all those chief groups of brooches which were to be found in Britain between the earlier part of the first century B.C., some perhaps a little earlier, and the latter part of the first century A.D. The longest sequences are the Nauheim–Nauheim Derivative and Drahtfibel–Drahtfibel Derivative, the former being the larger. Also included is a group of brooches which are related but do not form any part of those sequences: they are the "etc." in Part 1.

The purely continental types, The Rosette and Langton Down, did not give rise to large families of British copies or Derivatives. However, the Aesica of Part 4 and the Durotrigan Type of Chapter 6, Part 3, are offspring.

The Colchester derives from a continental form, but those made on the continent were never common in Britain and the form quickly became native and was the fons et origo of the bulk of British brooches in the latter part of the first century and well into the second century.

The Aesica is purely native owing only its general form to the Rosette: there is no slavish copying, principally because the methods of manufacture were different. The Glastonbury is a blip on the horizon: it appears to be purely British, but owes a lot to earlier forms of La Tène types.

The brooches of the King Harry Lane cemetery [675] are scattered throughout this chapter. As this is a very important and dated site, they occur in the sections on dating where only the phase in the excavation report is given: the dating of these phases is discussed in Appendix 1. These brooch types stand at the very beginning of the La Tène III series and have a spring system differing from earlier types whose integral springs coiled to one side only of the central axis of the bow. The novelty of the new types was that the spring was bilateral; it was evenly divided on either side of the axis, the two parts being joined by a chord. Almost invariably the initial coils were on the left-hand side when the brooch is looked at from the front, the chord crossing from the extreme left to the extreme right-hand side before the spring coils back to the centre. Therefore, the open face of the catch-plate was also on the right.

La Tène I brooches have a foot which generally turned up at an angle to the axis of the pin. The La Tène II brooch has the foot returned and tied to the bow either by a collar or by a wrap-round. The La Tène III family has the catch-plate as an integral part of the bow, there being no turned up foot. The collar, or wrap-round, is subsumed into the bow, but often survives as an ornamental feature. The last change marking the break between the La Tène II and the La Tène III is related to the spring. The La Tène II chord was always external and untrammelled. With the La Tène III, the chord was moved so that it lay at right angles to the pin, and fairly quickly became internal, i.e. running under the head of the bow. Alternatively, as we shall see in Part 3, it was held by a forward-facing hook.

This is a brief summary, and the fact that Britain produced several varieties which used hinged pins does not detract from the overall development, any more than the Involute Series ([385], Type 2Cb) which reverses the curve of the bow and can have a hinged pin, destroys what has been said. The three chief types forming Part 1 of this chapter all begin with external chords and end with internal ones and arguments that such a development may not be allowed to influence chronology are, in the writer's view, misplaced. The last section of the chapter is devoted to an aberration which rightly belongs here, but whose origins and influence need to be discussed.

Part 1. The Stead, Birdlip, Nauheim and Drahtfibel Group, etc.

[Late La T, N, ND, D, DD]

The Stead [Late La T]

Ian Stead [668], in studying the Iron Age, looked for something to act as a kind of marker for a horizon amongst the objects belonging to the Later pre-Roman Iron Age and chose brooches. His discussion sought to bypass the refined discussion of isolated rare items decorated in a distinctive style by seeking the more common found in association with other objects. However, ordinary archaeological sites so rarely yield such associations that brooches had hitherto been seen to be relatively useless. His was the first useful attempt to inject sense into this body of material whose examples, while never common, are frequent enough to serve as a class of object capable of useful sub-division. His study helped to define both the earlier phases and the end of the period. Since his paper, both the King Harry Lane [675] and the Westhampnett cemeteries [277] have been published.

I have, of course, not followed Dr. Stead absolutely, but used his basic criteria and have also abandoned the use of belgic, to define an overall type probably covering more elements than he would then have allowed. Stead was unwilling to accept that the external chord, a typologically early feature, had any real value in dating terms ([668], 410). A problem in dealing with a large class of object, such as brooches, is that a large number of sub-divisions needs some form of designation to render part of the whole mass intelligible. Numbers beyond a certain point are not easily memorised, hence the tendency to name a chief type after what is taken to be a type-site, but in the present case, not one is convenient for the whole. To call it Imitation La Tène II is too clumsy, even if the collar was important. Therefore, as Dr. Stead was the first to deal with it in this country, I decided to attach his name to it. Perhaps he will take it as a compliment to his contribution to Iron Age studies in Britain and beyond.

An interesting feature which