In Egyptian, Greek, and Roman decorative art the only kind of interlaced-work is the plait, without any modification whatever; and the man who discovered how to devise new patterns from a simple plait by making what I term _breaks_ laid the foundation of all the wonderfully complicated and truly bewildering forms of interlaced ornament found in such a masterpiece of the art of illumination as the Book of Kells in the library of Trinity College, Dublin. Although we do not know _who_ made this discovery of how to make breaks in a plait, we know pretty nearly _when_ it was made. In the decoration of the mosaic pavements in Great Britain belonging to the period of the Roman occupation, no instance, as far as I can ascertain, exists of the introduction of a break in a plait; nor is there any break in the plaitwork on the marble screen and the capitals of the columns of the ciborium in the Church of San Clemente at Rome (which are dated by R. Cattaneo[446] between A.D. 514 and 523). In the eighth century, however, there are several examples with well-authenticated dates of the use of true knotwork (as distinguished from plaitwork) in the decoration of churches in Italy; namely, on the ciborium of San Giorgio at Valpolicella[447] (A.D. 712); on the Baptistery of Calistus at Cividale[448] (A.D. 737); and on the jambs of the doorway of the Chapel of San Zeno in the Church of San Prassede at Rome[449] (A.D. 772-795). [446] _L’Architettura in Italia_, pp. 29 and 31. [447] _Ibid._, p. 80. [448] _Ibid._, p. 87. [449] _Archæologia_, vol. xl., p. 191. It would appear, then, that the transition from plaitwork to knotwork took place between the Lombard conquest of Italy under Alboin in A.D. 563, and the extinction of the Lombard monarchy by Charlemagne in A.D. 774; possibly during the reigns of Luitprand (A.D. 712-736) and Rachis (A.D. 744): for the name of the former king is mentioned in the inscriptions on the Baptistery at Cividale and the ciborium of San Giorgio at Valpolicella, and the latter on the altar at Cividale. [Illustration: PLAITWORK ON ROMANO-BRITISH PAVEMENT AT LYDNEY PARK, GLOUCESTERSHIRE (SIXTH CENTURY)] [Illustration: PLAITWORK ON CIBORIUM IN THE CHURCH OF SAN CLEMENTE, ROME] I now propose to explain how plaitwork is set out, and the method of making breaks in it. When it is required to fill in a rectangular panel with a plait the four sides of the panel are divided up into equal parts (except at the ends, where half a division is left), and the points thus found are joined, so as to form a network of diagonal lines. The plait is then drawn over these lines, in the manner shown on the accompanying diagram. The setting-out lines ought really to be double so as to define the width of the band composing the plait, but they are drawn single on the diagram in order to simplify the explanation. [Illustration: Regular plaitwork without any break] [Illustration: Method of making breaks in plaitwork]