Unstable editions of Ortelius' atlas

by Marcel P.R. van den Broecke

The author is a scientific adviser and Managing Director of the Company Cartographica Neerlandica in Bilthoven, the Netherlands which specialises in maps of Ortelius.

Portrait of Ortelius

Abraham Ortelius as depicted by Paul Rubens (1577-1640). (By courtesy of the Plantijn-Moretus Museum of Antwerp, where the original is displayed).

Abraham Ortelius'monumental work Theatrum Orbis Terrarum is regarded as the first atlas to appear. An atlas is defined in this instance as a uniform collection of map sheets of similar size, with sustaining text, compiled for the purpose of binding the sheets together to form a coherent book.

The Theatrum was an instant success and four issues of the first edition were published in 1570. When it appeared, it was the most expensive book ever printed. Despite this it was received by the public with such enthusiasm that no less than 7300 copies were produced in thirty-one editions from 1570 to 1612.1 Ortelius also issued about 750 copies of Additamenta and about 600 copies of the Parergon (maps from the antiquity), some separately, some bound up with the regular atlas to order. Ortelius is often characterised as being merely a publisher and compiler rather than a cartographical innovator but recent research by Peter Meurer2 has shown that the innovative nature of the form and content of the Theatrum should not be underestimated.

At least 900 copies are known to have survived to the present day,1 and there may be more because libraries, particularly in Eastern Europe, are still finding previously unrecorded copies. To obtain a better view on the question of how issues of the Theatrum relate to what is commonly regarded as an edition of an atlas I examined different copies of one edition and this research convinced me of the necessity to introduce the concept of unstable editions.

Throughout the various editions of Theatrum there are 228 different plates, 174 in the regular atlas editions and 55 in the Parergon. In spite of numerous attempts over the past century to establish a definitive list of plates, new ones are still being identified: since publication of Peter Meurer's book2 in 1991 and Robert Karrow's book in 19933, Meurer has found plate 71/II 'Hannonia' 1575, and I have found 93/I 'Americae' 1579-84 (Meurer's 93/I 'Asiae'should be renumbered 71/I as this plate does not stem from 1579 but from 1575), 93/IV 'Valentiae', 116/I 'Angliae, Scotiae et Hiberniae' 1589, 135/VI 'Palestinae' 1595-1612, 33P/II 'Italiae Veteris' 1595-1624 and 39P/IV 'Erythraeae' 1609-1624. The first edition of 1570 contains 53 plates. the largest edition of Theatrum, the English edition of 1606, contains as many as 166 of the total 228 plates; 29 are from the first edition. Clearly, many new plates were introduced and only a few cast aside. An up-to-date inventory of plates is given further on in this article.

Two states of the cartouche of the Flandriae plate,occurring in two exemplars (Leiden University and Amsterdam University) of the first version of the first edition of the Theatrum, (Koeman's Ort 1A). The Leiden exemplar shows a heart, the Amsterdam exemplar does not.

An "unintended" state of a map in the form of plate damage can also help in determining its age. In this case plate damage consists of a crack which began to form in the cartouche of the continent map of Africa after 1602 and which progressively widened with each subsequent edition. This is what the cartouche looks like in the editions of 1612.

As a true humanist and renaissance cartographer, Ortelius was keenly interested in the geographical knowledge of classical antiquity. The Parergon maps, which he considered his major cartographical achievement, bear witness to this. As early as 1578, Ortelius knew about the existence of the Peutinger tables and tried to get hold of them. They show the Roman world view around the third century. The original, found by Konrad Celtes (1459-1508) in a library in Augsburg, came into the hands of Konrad Peutinger (1465-1547) and later went to his relative Marcus Welser. Welser was the first to publish a copy of it in 1591 at Aldus Manutius in Venice, which Ortelius possessed but found inadequate and insufficient, since it seemed to deviate in many respects from the original. In 1598 new, more accurate copies were made by Welser at Ortelius'request. These formed the basis of his Peutinger tables in four sheets. Proof prints from these plates went back to Marcus, were compared with the original and corrected accordingly. The original Peutinger tables disappeared, were found back in 1714 and are now in theNational Library in Vienna. Because of damage and progressive blackening of colours of the 11 (once 12) original sheets of parchment, together once forming a roll, Ortelius' version now is the most reliable representation. Ortelius supervised the engravings but did not live to see the final result, which was first published by Moretus in 1598 as a separate booklet with text by Ortelius. Subsequently, Bertius included prints from these plates in his Theatrum Geographiae Veteris without text. It was not until 1624 that the tables finally appeared in the last edition of Ortelius' Parergon, produced by Plantijn-Moretus. To demonstrate the importance Ortelius attached to these tables, consider his final piece of text accompanying them (translated from Latin): "Farewell dear spectator and dear reader, enjoy this monument which, although it has plenty of shortcomings, does not have an equal or even anything like it under all the relicts from antiquity".

The sheets form a road map. The North-South dimsension is heavily crushed. The first segment begins with England and Spain. Most space and greatest accuracy is devoted to Italy and Greece (segments 2-5). The Roman world ends in the east with the river Ganges and island of Ceylon. Shown here are the first and last segment.

The first and most obvious reason for discarding a plate and replacing it with a new one is increased geographical knowledge. Knowledge of the world expanded at a tremendous pace during the life cycle of the Theatrum and examples where incorrect information leads to correction by introducing a new plate are plentiful. The bulge on the west coast of South America on plates 1 and 2, corrected on plates 113 and 114, are the best known examples but there are others, even among the Parergon plates.1

The second reason for introducing a new plate also goes back to increased knowledge. Whereas Ortelius had insufficient information (and perhaps insufficient economical resources) to include a map of the Pacific or Japan for example, the commercial succes of the atlas and the information he obtained from new sources warranted later inclusion. Uncertainty about the reliability of existing maps also led him in various instances to add another map of the same region, often with different information. This allowed the user to choose which map was more pleasing. The addition of a second map of Hungary (91) next to the exisitng one (42) is an example of this, as both plates continued to be altered during their life-cyrcle.

A third reason for introducing a new plate was wear of the old one. The second plate of the world map (112/I), replacing plate 1, still with the bulge in the South American coastline, is a case in point. From 1575 onwards, plate 1 developed a crack in its lower left corner and the bolts applied to keep it together were only a temporary solution.

Finally, there is a mystery category. It is unclear why a new Abraham plate (25P/I) was made next to the old one (12P) since they were used side by side for some time. They differ mainly in the diagonal direction of the background hatching and must have taken a great deal of time and effort to engrave due to the delicacy of the scenes in the twenty-two medaillions surrounding the map. Maybe the plate was lost for a time.

Then there is the matter of copyright or privilegio. An older version of 'Artois' (map 82) was reintroduced at a later stage, when a newer version (map 115) was already available. This had to do with privilege as indicated by Denucé4. Some offprints of this map were included in the Latin edition of 1575 in the expectation that privilige would be obtained from Philip II but when it was refused a new version was made. When the privilege was finally granted, the old plate reappeared. Finally, as Meurer points out, there may also have been parallel printing of nearly identical plates on different presses to speed up production. Homann at one stage employed three plates for one map in parallel. These prints can be distinguished only by differing widths between the edge of the print to the edge of the platemark.

No systematic, exhaustive attempt has so far been made to identify the various states of each of the plates and it is beyond the scope of this paper. However, on the basis of meticulous inspection of various copies from a few plates in different editions (such as plates 41, 42 and 43 by László Gróf5), plate 18 'Zeelandicarum' by Frans Gittenberger (unpublished), plates 1, 112/I, and 113 'Typus Orbis'6 or plate 19 'Hollandiae' in my paper1, it seems fair to assume that all plates except the very shortlived ones existed in more than one state. This is understandable as copperplates wore out quickly and needed recutting after 1000 impressions. Some had even shorter lives giving only 300 impressions. Correcting a plate was obviously more economical than replacing it.

In addition, there are also different variants which resulted from damage to the plate (see letter from R. Shirley to the Editor of TMC about definition of state. Issue 67, p.56). The first plate of the world map is a case in point; in 1575 its lower left corner broke off. An engraved copperplate took two to six months to engrave it represented a considerable investment and could not be easily discarded. Therefore, copperplates were repaired whenever that turned out to be feasible. The world map plate was repaired by bolting a sustaining piece of copper to the back of the plate. An image of these bolts can be seen in the impresiions and help identify the version. Shirley7 and personal communication discern the following states for the three plates of the Ortelius world map:

1 Typus Orbis Terrarum, plate 1

State 1 1570-1579 (from 1575 with bolt impressions in lower left corner)

State 2 1579-1584 crack somewhat mended, clouds reworked

State 3 1584-1585 date 1584 (or 1585) added to the right of Franciscus Hogenberg Sculpsit

112/I Typus Orbis Terrarum, plate 2

State 1 1586 unsigned, slightly smaller than plate 1, bordering clouds retained, dated 1586

State 2 1587-1589 as State 1 but without date

State 3 1588-1589 as State 2 but without corrected coastline of South America

113 Typus Orbis Terrarum , plate 3

State 1 1588-1612 medallions in corners, geographically revised, dated 1587

State 2 Le Maire Strait added, date removed

Another case of damage through use, this time with no repair attempted, is the plate of the African continent. From 1602 onwards the part of the copperplate for the lettering Africae Tabula Nova in the main cartouche began to crack. The later the print, the clearer these cracks become. Obviously, such unintended 'states' can be used to date maps just as succesfully as purposeful alterations although they are not formally recognised as new states.

Clearly, states constitute an undispensable tool for dating maps. Such information allows loose maps to be linked to specific editions. However, this sounds simpler than it turns out to be. it is my firm conviction that a more meticulous and systematic analysis than the customary informal analysis by eye (i.e. perhaps by computerised flatbed image scanning) would reveal many more states of each plate than have been identified so far.

Rather than showing wat the Romans knew of the world, this map shows the Roman empire itself. The medallions in the top corners show the legendary founders Romulus and Remus, the three at the lower right corner the lineage of the subsequent Roman kings (double circles) and their wives (single circles), based on thi historical writings of Livius, Dionysius and Plutarchus.

This map shows the area where Alexander the Great made his conquests, beginning in Egypt, where the Ammon-Jupiter oracle (see lower left) welcomed him as "Son of Zeus" and predicted a great future for him. The numerous cities in Persia called Alexandria which he founded are also shown.

Regular atlas maps



period of usage

numbers printed


1. Typus Orbis Terrarum 1570L–1584L 3250  
2. Americae 1570L–1575L 1675  
3. Asiae 1570L–1574L 1575  
4. Africae 1570L–1612S 7300  
5. Europae 1570L–1581F 3025  
6. Angliae, Scotiae & Hiberniae 1570L–1612S 7300  
7. Regni Hispaniae 1570L–1612S 7300  
8. Portugalliae 1570L–1612S 7300  
9. Galliae 1570L–1603L 5650  
10. Biturgum-Limaniae 1570L–1612S 7200 not in 1598D
11. Caletensum-Veromandorum 1570L–1598D 4450  
12. Galliae-Narbonensis-Sabaudiae 1570L–1581F 3025  
13. Germaniae 1570L–1602G 5350  
14. Germaniae Inferioris 1570L–1606E 5950  
15. Gelriae, Cliviae 1570L–1612S 7300 not in 1592L and some in 1595L
16. Brabantiae 1570L–1592L 4150  
17. Flandriae 1570L–1575L 1675  
18. Zeelandicarum 1570L–1612S 7300  
19. Hollandiae Catthorum 1570L–1612S 7300  
20. Oost en West Friesland 1570L–1612S 7300  
21. Daniae Regnum 1570L–1581F 3025  
22. Thietmarsiae-Prussiae 1570L–1581F 3025  
23. Saxoniae-Misniae-Thrungiae 1570L–1612S 7300  
24. Franconiae-Osnabrugensis 1570L–1612S 7300  
25. Regni Bohemiae 1570L–1612S 4150  
26. Silesiae 1570L–1592L 4150  
27. Austriae 1570L–1592L 4250 also in 1598D used side by side with 135/IV in 1595L
28. Salisburgensis 1570L–1595L 4250  
29. Typus Vindelicae 1570L–1573G 1250  
30. Bavariae-Wirtembergensis 1570L–1581F 2875  
31. Helvetiae 1570L–1612S 7300  
32. Italiae 1570L–1581F 2875  
33. Ducatus Mediolanensis 1570L–1612S 7300  
34. Pedemontana 1570L–1612S 7300  
35. Como-Romae-Friuli 1570L–1612S 7300  
36. Thusciae 1570L–1612S 7300  
37. Regni Napolitanie 1570L–1612S 7300  
38. Insularum Aliquot 1570L–1612S 7300  
39. Cyprus-Candia 1570L–1581F 2875  
40. Graeciae 1570L–1612S 7300  
41. Sclavoniae-Croatiae-Carniae 1570L–1612S 7225 not in 1573D
42. Hungariae 1570L–1612S 7300  
43. Transylvaniae 1570L–1575L 1675  
44. Poloniae 1570L–1595L 4250 used side by side with 135/IV in 1595L
45. Septentrionalium Regionum 1570L–1612S 7300  
46. Russiae 1570L–1612S 7300  
47. Tartariae 1570L–1612S 7300  
48. Indiae Orientalis 1570L–1612S 7025 not in 1589F
49. Persici 1570L–1612S 7300  
50. Turcii Imperii 1570L–1575L 1675  
51. Palestinae 1570L–1575L 1675  
52. Natoliae-Aegypti-Cartaginis 1570L–1612S 7300  
53. Barbariae 1570L–1612S 7300  
53/I. Hannoniae never regularly in atlas

5 copies known with date 1572, one without date

54. Scotiae 1573GI–1612S 6275  
55. Angliae 1573GI–1602L 4625 also in 1606E
56. Cambriae Typus 1573GI–1612S 6275  
57. Hiberniae 1573GI–1603L 3925  
58. Mansfeldiae 1573GI–1612S 5850  
59. Thuringiae-Misniae 1573GI–1612S 6025  
60. Moraviae 1573GI–1612S 6025  
61. Basilensis-Suaviae 1573GI–1612S 6175 not in 1598D
62. Rhetiae-Goritiae 1573GI–1612S 6175 not in 1598D
63. Fori Iulii 1573GI–1612S 6175 not in 1598D
64. Patavini-Apuliae 1573GI–1592L 3825 also in 1598D, 1598F
65. Senensis-Corsica-Ancona 1573GI–1612S 6275  
66. Cypri Insulae 1573GI–1612S 6275  
67. Carinthiae-Histriae-Zarae 1573GI–1612S 6275  
68. Pomeraniae-Livoniae-Oswiec. 1573GI–1612S 6275  
69. Presbiteri Johannis 1573GI–1612S 6275  
70. Bavariae 1573LI–1612S 6100  
71. Illyricum 1573LI–1612S 6000 not in 1598D
71/I. Asiae 1575L–1612S 5625  
72. Hispaniae Nova 1579LII–1612S 5875  
73. Culicaniae, Hispaniolae, Cub. 1579LII–1612S 5875  
74. Hispalensis Conventus 1579LII–1612S 5875  
75. Pictonum 1579LII–1612S 5875  
76. Anjou 1579LII–1612S 5775 not in 1598D
77. Picardiae 1579LII–1612S 5875  
78. Burgundiae 1579LII–1612S 4275 not in 1589G, 1598D, 1603L, 1612I, 1612S
79. Lutzemburgensis 1579LII–1612S 5875  
80. Namurcum 1579LII–1612S 5875  
81. Hannoniae 1579LII–1581F 1450  
82. Artois 1579LII–1584F 1725  
83. Frisia Occidentalis 1579LII–1612S 5875  
84. Frisiae Orientalis 1579LII–1592L 2725  
85. Westphaliae 1579LII–1612S 5875  
86. Hassiae-Holsatiae 1579LII–1592L 2725 also in 1598F
87. Burghaviae-Waldeccensis 1579LII–1612S 5775 not in 1598D
88. Wirtenberg Ducatus 1579LII–1612S 5775 not in 1598D
89. Veronae Urbis 1579LII–1612S 5775 not in 1598D
90. Agri Ceremonensis 1579LII–1612S 5775 not in 1598D
91. Ungariae Loca 1579LII–1612S 5775 not in 1598D
92. Turcii Imperii 1579L–1612S 5625  
93. Palestinae 1579L–1592L 2725  
93/I. Americae 1589L–1584L 1575  
93/II. Flandria 1579L–1589G 2375 also in some 1592L
93/III. Transilvania 1579L–1612S 5625  
93/IV. Valentiae none   only one copy known
94 . Bavariae-Argentoratensis 1584LIII–1612S 4425  
95. Acores 1584LIII–1612S 4425  
96. Burgundiae Inferioris 1584LIII–1612S 4075 not in 1598D, 1602S
97. Chinae 1584LIII–1612S 4425  
98. Candia-Archipelagi 1584LIII–1612S 4425  
99. Carpetaniae-Guipuscoae 1584LIII–1612S 4425  
100. Leodiensis 1584LIII–1595L 1575  
101. Daniae-Oldenburg 1584LIII–1612S 3975 also with 109, 130
102. Perusini Agri 1584LIII–1612S 4325 not in 1598D
103. Peruviae-Florida-Guastecan 1584LIII–1612S 4425  
104. Prussiae 1584LIII–1595L 1575  
105. Romaniae 1584LIII–1612S 4325 not in 1598D
106. Terra Sancta 1584LIII–1612S 4425  
107. Valentiae 1584LIII–1612S 4325 not in 1598D
108. Gallia Narb.-Sabaud.-Venux. 1584LIII–1612S 4425  
109. Holsatiae-Rugiae 1584LIII–1612S 4425 also with 101, 86
110. Europae 1584L–1612S 4425  
111. Hannoniae 1584L–1612S 4175 not in 1602G
112. Italiae 1584L–1612S 4425  
112/I. Typus Orbis 1588S 300 used side by side with 113, 1587–1589
113. Typus Orbis 1587F–1612S 3850  
114. Americae 1587F–1612S 4150  
115. Artesia 1587F–1612S 3800 not in 1598D, 1602G
116. Burgundia Comitatus 1589G & 1603L 450  
116/I. Angliae, Scotiae et Hiberniae 1589 composite few noted by A. Kelly
117. Aprutti 1590LIV–1612S 3450 not in 1598D
118. Brandenburgensis 1590LIV–1612S 3550  
119. Bresciano 1590LIV–1612S 3450 not in 1598D
120. Ischia 1590LIV–1612S 3550  
121. Islandia 1590LIV–1612S 3200 not in 1598F, 1602G
122. Lotharingiae 1590LIV–1612S 3550  
123. Braunswicensis-Norimbergae 1590LIV–1612S 3550  
124. Maris Pacifici 1590LIV–1612S 3550  
125. Flandria 1592L–1612S 3450  
126. Brabantiae 1592L–1612S 3450  
126/I. Gelriae Cliviae 1592L 300 also in some 1595L
127. Cenomanorum-Neustria 1595LV–1612S 3050 not in 1598F, 1598D
128. Provinciae 1595LV–1612S 3150 not in 1598D
129. Hennebergensis-Hassiae 1595LV–1612S 3050 not in 1598F, 1598D also with 86
130. Daniae-Cimbriae 1595LV–1612S 3150 not in 1598D, also with 101
131. Patavini-Tarvisini 1595LV–1612S 3050 not in 1598F, 1598D
132. Florentini 1595LV–1612S 3150 not in 1598D
133. Apuliae-Calabriae 1595LV–1612S 3050 not in 1598F, 1598D
134. Iaponiae 1595LV–1612S 3250  
135. Fessae et Marocchi 1595LV–1612S 3250  
135/I. Frisiae Orientalis 1595L–1612S 3150  
135/II Silesiae 1595L–1612S 3150  
135/III. Gallia none 2? never regularly included, cartouche from Mercator, Zeelandica
135/IV. Poloniae Lithunaiae 1595L–1612S 3050 used side by side with 44 in 1595L
135/V. Salisburgensis 1595L–1612S 3050 used side by side with 28 in 1595L
135/VI. Palestinae 1595L–1612S 2550 not in 1598F, 1602S, 1602G
136. Isle de France 1598F–1612S 2850  
137. Tourraine 1598F–1612S 2850  
138. Blaisois-Lemovicum 1598F–1612S 2850  
139. Caletensium-Veromanduorum 1598F–1612S 2850  
140. Austriae 1598F–1612S 3050 not in 1598D
141. Prussiae 1598F–1612S 2850  
141/I. Leodiensis 1598F–1612S 2850  
142. Burgundiae Ducatus-Comitatus 1602S–1612S 1300 not in 1606E, 1608I
143. Angliae Regnum 1603L 300 also in some 1602G, 1602S, 1608I, 1609L, 1610D
144. Cataloniae Principatus 1603L–1612S 1350 not in 1606E
145. Reyno de Galizia 1598D–1612S 1750  
146. 116   2050  
147. Deutschland 1603L–1612S 1650  
148. Angliae et Hiberniae 1606E–1612S 1650  
149. Irlandiae 1606E–1612S 1650  
150. Gallia 1606E–1612S 1650  
151. Limburgensis Ducatus 1606E–1612S 1650  
152. Lacus Lemani 1608I–1612S 1050  
153. Inferioris Germaniae 1608I–1612S 1050  
154. Bononiensis-Vicentini 1608I–1612S 1050  
155. Reipublicae Genuensis 1608I–1612S 1050  
156. Parmae et Placentiae 1608I–1612S 1050  
157. Ducatus Ferrariensis 1608I–1612S 1050  
158. Romagna olim Flaminia 1608I–1612S 1050  
159. Ducatus Urbini 1608I–1612S 1050  
All regular atlas maps (174 plates) about 730.000
Parergon maps
1P. Divi Pauli 1579LII–1624P 5875  
2P. Romani Imperium 1579LII–1624P 5875  
3P. Hellas, Graeciae Sophiani 1579LII–1624P 5625 not in 1608I
4P. Aegyptus (North) 1584LIII–1592L 975  
5P. Aegyptus (South) 1584LIII–1592L 975  
6P. Belgii Veteris Typus 1584LIII–1592L 975  
7P. Creta-Ins. Aliquot 1584LIII–1624P 4275 not in 1598D
8P. Insularum Aliquot-Cyprus 1584LIII–1624P 4275 not in 1598D
9P. Italiae Veteris Specimen 1584LIII–1592L 975  
10P. Siciliae Veteris Typus 1584LIII–1624P 4275 not in 1598D
11P. Tusciae Antiqua Typus 1584LIII–1592L 975  
12P. Abrahami Patriarchae 1590LIV–1595L 400 used side by side with 25P/I, 1592–1595
13P. Aevi Veteris Typus 1590LIV–1624P 3800  
14P. Africae Propriae Tabula 1590LIV–1624P 3700 not in 1598D
15P. Brittanicae Insularum (North) 1590LIV–1592L 400  
16P. Brittanicae Insularum (South) 1590LIV–1592L 400  
17P. Pontus Euxinus 1590LIV–1624P 3700 not in 1598D
18P. Gallia Vetus 1590LIV–1602G 1150 not in 1598D
19P. Germaniae Veteris Typus 1590LIV–1624P 3700 not in 1598D
20P. Hispaniae veteris Descriptio 1590LIV–1624P 3700 not in 1598D
21P. Italia Gallica Cisalpina 1590LIV–1624P 3700 not in 1598D
22P. Typus Chorographicus 1590LIV–1624P 3700 not in 1598D
23P. Pannoniae et Illyrici 1590LIV–1624P 3700 not in 1598D
24P. Tempe 1590LIV–1624P 3800  
25P. Thraciae Veteris Typus 1590LIV–1624P 3700 not in 1598D
25P/I Abrahami Patriarchae 1592L–1624P 3200 used side by side with 12P, 1592–1595
26P. Europam sive Celticam 1595L–1624P 3400 not in 1598D
27P. Galliae Veteris Typus 1595L–1624P 3400 not in 1598D
28P. Latium 1595L–1624P 3400 not in 1598D
29P. Graecia Major 1595L–1624P 3400 not in 1598D
30P. Daciarum Moesiarumque 1595L–1624P 3400 not in 1598D
31P. Alexandri Magni 1595L–1624P 3400 not in 1598D
32P. Aeneas Troiani Navigatio 1595LV–1602G 1250 not in 1598D
33P. Daphne 1595LV–1624P 3400 not in 1598D
33P/I. Belgii Veteris Typus 1595L–1624P 3300 not in 1598D
33P/II. Italiae Veteris Specimen 1595L–1624P 3300 not in 1598D
33P/III. Tusciae Antiquae Typus 1595L–1624P 3300 not in 1598D
34P. Brittanicarum Insularum 1595L–1624P 3300 not in 1598D
35P. Aegyptus Antiqua 1595L–1624P 3300 not in 1598D
36P. Geographia Sacra 1598F–1603L 950  
37P. Erythraeae sive Rubri Maris 1598F–1608I 1550  
38P. Argonautica 1698F–1624P 3150  
39P. Scenographia Escoriali 1601L–1624P 3050  
39PI. Gallia Vetus 1603L–1624P 2550  
39PII. Aeneas Troiani Navigatio 1603L–1624P 2550  
39PIII. Geographia Sacra 1606E–1624P 1950  
39PIV. Erythraeae sive Rubri Maris 1609L–1624P 1350  
40P. Ordines Sacri I 1603L–1624P 2550  
41P. Ordines Sacri II 1603L–1624P 2550  
42P. Lumen Orientalis 1624P 300  
43P. Lumen Occidentalis 1624P 300  
44P. Peutingerorum(I,II) 1624P 300 separately issued 1598
45P. Peutingerorum(III, IV) 1624P 300 separately issued 1598
46P. Peutingerorum (V, VI) 1624P 300 separately issued 1598
47P. Peutingerorum (VII, VIII) 1624P 300 separately issued 1598
All Parergon maps (55 plates) about 143.000
Total of all maps (228 plates) about 873.000


But what about the development of texts in the Theatrum? Since most editions contain texts in Latin, which is no longer the language of science, this aspect has been neglected. Even Meurer has confined his search of sources for each plate to the Catalogus Auctorum2 and has not analysed the development of the text accompanying each plate in systematic detail. A text history, in some cases augmented by information on the paper watermark, is useful in dating maps. Texts and their development over various editions also provide insight into new information, new sources, and the nature of the feedback that Ortelius received from his customers and fellow cartographers.

Descriptions of various atlas editions in cartobibliographical literature such as Koeman's Atlantes Neerlandici8 raises the expectation that all exemplars of that edition are identical in texts as well as plates. However, inspection of two copies of what Koeman calls edition 1(A) of the Theatrum in the University Libraries of Amsterdam and Leiden 1(B) does not confirm this. The maps of "Flandria" turn out to derive from two different parallel plates, as also pointed out by Gittenberger: the cartouches differ at the middle in top and bottom area (see illustration)9. The Leiden exemplar shows a heart, whereas that of Amsterdam does not. Comparison of the texts between the Leiden exemplar and the Amsterdam exemplar yield considerable differences: in (A) the title of map 5 (Europe) is reset when compared with (B). In plate 8 (Portugalliae), 9 (Galliae), 34 (Pedemontanae), 35 (Como), 46 (Russiae), 47 (Tartariae) and 51 (Palestinae) the whole text is reset. Sometimes it is longer in (A), sometimes longer in (B).

How can this be and which of these exemplars from supposedly the same variant of the same edition came first? It is tempting to regard the version with the most text as the later one, as texts tend to grow in size over editions, but even this simple question cannot be answered as the first exemplar has more text in some cases than the later exemplar. Are these differences exceptional and restricted to this variant of the first edition, or is it the rule? In order to answer this question I decided to campare three exemplars of the 1595 Latin edition, the last one that Ortelius himselfproduced before his death in 1598. These are atlas 1803 A7 in Amsterdam University Library, a copy from the Florence Military Library that has appeared in facsimile10 and atlas 1802 A6 from Amsterdam University Library.

Differences between all three exemplars are numerous: for 1803 A7 left-over sheets dating back to the 1584 Latin and even to the 1575 Latin edition were used. Text sheets differ not only in type and length but also in order. The order of the Parergon maps in the Florence exemplar is erratic in spite of its old binding. Differences in the maps and texts between exemplars of one edition turn out to be the rule rather than the exception.

Of all exemplars (about ninety) of the Theatrum offered for sale and described in detail by the major auction houses between 1980 and 1992 almost half do not conform to the Koeman descriptions. They may contain maps or a nomenclature not called for, or lack one that should have been included, display other text sections which are inconsistently included or excluded, or contain Additamenta included in an unpredictable manner (one 1570 Latin edition offered recently contains two identical 1574 Latin I Additamenta bound up with it) or editions may contain maps from de Jode's Seculum Orbis Terrarum or Braun and Hogenberg's Civitates Orbis Terrarum etc.

A typical example to further demonstrate variability between exemplars of the same edition is a composite atlas with a late binding acquired at Sotheby's, London, on April 23, 1987; it contains some maps from de Jode's Speculum … and a few manuscript maps inserted at the appropriate places. It also contains, in mutilated form, the mysterious 71/II Hannonia map, mentioned in Denucé4 (Volume I p. 36 and Volume II p.41 ff) which never appeared regularly in Theatrum editions due to lack of privilegio (cf. Meurer11). Title-page and colophon qualify this atlas as a 1573 Latin exemplar, but it also has maps with page numbers from the 1573 Latin I Additamentum inserted at the appropriate places. Moreover, it contains both the early Bavaria map 29: "Typus Vindelicae" (1570 Latin-1573 German edition) and the later map 70: Bavaria (1573 Latin Additamentum I-1612 Spanish edition).

It seems that the notion "edition" needs a redefinition in view of this data. When we buy a book today we expect to obtain a physical object which is identical with the same edition of that book bought by everyone else. But clearly this has not always been the case. For early atlases like the Theatrum the definition of unstable editions is needed. This becomes better understood when the production methods of thge atlas are examined.

Maps were printed intaglio from copperplates on which the information was engraved by mirror image. Plates were inked, then wiped clean and the ink remaining in the copper grooves was transferred to the damp paper by applying high pressure from a roller. The resulting prints or pulls were hung up to dry. Texts were composed by setting individual letters in reverse order into rows forming the printed lines, and putting these one beneath another until the text of an entire page had been set. These rows of letters were then fixed on a page block, inked and block pressed on to the paper at relatively low pressure. In principle there is no reson why either text or map should be printed first but in practice there is a big difference: type could be removed from the page block after use and re-used for composing new texts whereas a copperplate is a more fixed, permanent medium. Individual letters used for typesetting represented a considerable investment for the printer and were always in high demand and short supply which is the reason it was not feasible to typeset an atlas in one go. Instead, typeset pages were broken up as soon as enough copies had been drawn from it in order to re-use the type for subsequent pages of text. Therefore, it was in the printer's interest to print the text sheets in sufficient numbers, and only to print the maps on verso as needed. An exception to this is the 1606 English edition.

The red sea, which in Ortelius rendering of classical knowledge extends along both sides of the Indian peninsula, does not derive its name from being red, but from Perseus' son Erythras (the red king) who was buried on some island in this sea. Note the relatively accurate representation of the Indian continent and the IndoChinese/Malaccan peninsula. The inset map shows the various places that Ulysses visited on his wanderings from Troy to his native Ithaca, as first described in Homer's Odysse. This new plate (39P/IV) is different from its predecessor (37P) in being slightly larger, and in having more text and a different cartouche.

The chance of there being an exactly equal number of sheets in one print run is small, particularly as some pulls would fail in quality and be discarded. This means that a new edition may contain old sheets that were left over from a previous printing with old texts and old plates, or with old texts and new plates. To complicate matters further, it might also contain new texts and old plates; in spite of the arguments above it was not uncommon to print some map plates on sheets without text in order to sell them individually. Some 50% of the loose Ortelius maps in my possession are without text on verso. Once a new edition was being prepared, old stockpiles of such sheets might be provided with texts and included in the new edition for economic reasons.

It is unlikely, in view of this, that all exemplars of a single issue of the Theatrum are exactly alike as regards plates and texts. When two versions of a plate were available side by side it is still possible that a mix up could occur and that an earlier issue appeared with the later plate or vice versa. This is what happened with the 1592 Latin which has the new Abraham plate (25P/I) with the background hachuring between the medallions going from lowr right to upper left. Similarly plate 2 and 3 of the world map (112/I and 113) co-existed for some time. 112/I occurs in the 1588 Spanish edition but also in the 1589 German edition in its first and second state, with a bulge in South America. The bulge was corrected on this plate but the impressions not included regularly in atlases. Plate 113, in which the bulge has been corrected and which features medallions rather than clouds in the corners, occurs in the 1587 French and 1589 German editions and is from then on used as the only world plate. For text, too, variations may be expected to be the rule rather than the exception.

The physical characteristics of a printer's shop at this time must also be taken into account in this argument. Even Blaeu's wellknown premises at the Bloemgracht in Amsterdamn, a century after Ortelius, produced its momumental output from premises measuring only 160 sq. metres (approx. 1600 sq.feet). Ortelius' Latin edition of the Theatrum was produced from the famous Plantin printing presses in Antwerp along with many others.10 The 300 or so piles of map and text sheets, each at least 200 sheets high, would be waiting to be bound in cramped conditions. The chances of mistakes being made when these sheets were collated for binding were high and this is what we found. In addition, atlases were often left unbound until the buyer had been consulted about the binding required. All this contributes to the instability of an edition.

On the basis of our findings we do not agree with Koeman's premise that an atlas edition may be regarded as the standard by which to compare other copies. Each copy of these manually produced atlases is unique and each edition will display a certain degree of instability in respect to other exemplars of that edition. The degree of instability of each Theatrum edition and of other early atlases produced in a similar manner (Meurer reports a similar instability amongst atlases by Sebastian Münster) can only be determined by comparing many exemplars of each edition. Fortunately, enough exemplars of each edition of the Theatrum still exist in order for us to do this.


1 Marcel van den Broecke, "How rare is a map and the atlas it comes from? Facts and speculations on production and survival of Ortelius' Theatrum Orbis Terrarum and its maps" in The Map Collector 36, pp.2-12.

2 Peter H. Meurer, "Fontes Cartographici Orteliani: das 'Theatrum Orbis Terrarum' von Abraham Ortelius und seine Kartenquellen', (Weinheim: VCH, 1991).

3 Robert W. Karrow, Mapmakers of the sixteenth century and their maps (Winnetka, Chicago: Speculum Orbis Press, 1993).

4 Jan Denucé, Oud-Nederlandsche kaartmakers in betrekking met Plantijn, (Amsterdam: Meridian Publishing Company, 1912).

5 Làszló Gróf, "Magyarország térképei az Ortelius atlas zokban" in Cartographica Hungarica 1 (1992) pp.26036.

6 Rodney W. Shirley, Early printed maps of the British Isles 1477-1650 (East Grinstead: Antique Atlas, 1991).

7 Rodney W. Shirley, The Mapping of the World: early printed world maps 1472-1700 (London: Holland Press, 1993).

8 C. Koeman, Atlantes Neerlandici Vol. 3 (Amsterdam: Theatrum Orbis Terrarum, 1969).

9 Personal communication.

10 A. Ortelius, Theatrum Orbis Terrarum Facsimile of the 1595 edition (Firenze, Giunti, 1991).

11 Peter H. Meurer, "De verboden eerste uitgave van de Henegouwen-kaart door Jacques de Surhorn uit het jaar 1572 in Caert Thresoor 13 (3), 1994, pp. 81-87.


This paper, which will also appear in a shortened form in Dutch Caert Thresoor, has benefited from comments by C. Koeman, D. de Vries, B. Buijnsters, R. Shirley, A. Kelly, L. Grof, F. Gittenberg, P. Meurer, G. Ritzlin and R. Karrow.