¹ I wish to express my sincere thanks to Dr. N. H. Baynes for his most generous help and stimulating criticism, and to Professor R. J. H. Jenkins for his constant encouragement and innumerable valuable suggestions.

² A short bibliography of the more important works will be found in Baynes, N. H., The Byzantine Empire (Home University Library, Oxford University Press, 1944), 250 fGoogle Scholar. They are almost exclusively concerned with the Thematic organization. Three other works which touch on the more specialized field of naval architecture are: Torr, Cecil, Ancient Ships (Cambridge University Press, 1895)Google Scholar; Bury, J. B., Appendix to vol VI of his edition of Gibbon (London: Methuen, 1898)Google Scholar; Mueller, K. K., Eine griechische Schrift über Seekrieg (Würzburg, 1882)Google Scholar. To these should now be added W. L. Rodgers, Naval Warfare Under Oars 4th–6th Centuries, Annapolis, 1939, a thoughtful sketch written from a seaman's viewpoint.

³ Dain, Alphonse, Naumachica (Paris, 1943)Google Scholar.

⁴ All references to the Περὶ θαλασσομαχίος the ‘Anon PBPP’, the 'Εκ τῶν Τακτικῶν of Nicephorus Uranus and the Ναυμαχικά of Syrianus Magister are to the text of M. Dain. An inferior text of Leo, differently divided, is printed in MPG, CVII, as Constitutio XIX of the ‘Tactica’.

I hope to discuss elsewhere my reasons for dating the Περὶ θαλασσομαχίος so narrowly—the date seeming to depend on the chronology adopted for the years 904–11. The ‘Anon PBPP’ was first published by Fabricius, but was not reprinted in the later edition of his works. This latter fact probably explains why it has not figured in the great ‘trireme controversy’, although one passage at least is extremely pertinent. The date of its composition has been discussed at some length by Stilpon Kyriakides, Atti del V Congresso Internazionale di Studi Bizantini 501 f.

⁵ These inventories have been preserved with the text of the so-called De Caerimonnis attributed to Constantine Porphyrogenitus. In the Bonn edition they appear as chapters 44 and 45 of Book II.

⁶ I must here acknowledge the kindness of Mr. E. W. White, of the Science Museum, who discussed with me many points relating to the Western ships of this period.

⁷ I must apologize for the cursory simplification of many outstanding problems which are well worth discussion. Unfortunately they all entail reference to sources and detailed examination of quite a formidable literature.

⁸ The parallel use of the terms χελάδιον and δρόμων is a subject that merits the closest study.

⁹ The origin of the term πάμφυλος is debatable. Gfrorer saw a survival of the classical sense, while Bury suggested a Pamphylian derivation. I prefer to follow Gfrorer, and would meet Bury's objections by a different interpretation of Leo Περὶ θαλ. 42.

¹⁰ The principal sources for the above are Leo Περὶ θαλ. passim, and the inventories in Porphyrogenitus.

¹¹ Περὶ θαλ. 8.

¹² Torr, op. cit., 22; N. G. L. Hammond, JHS, 1945, 33, n. 25. The galleys of the Knights may best be studied in the rich collection of prints and models at the Science Museum, Kensington.

¹³ Theophanes Continuatus, Vita Bas. Mac., c. 61.

¹⁴ For the fifteen/two ratio see Torr, op. cit., 22–3.

¹⁵ ‘Anon PBPP,’ 2, 1.

¹⁶ σταμῑνες.

¹⁷ μήτραι and ἐγκοίλια, cf. the ‘matere’ of the later mediaeval galley, Torr, op. cit., 39, n. 95.

¹⁸ ‘Anon PBPP,’ 2, 11.

¹⁹ Περὶ θαλ. 69.

²⁰ The ἂφλαστα were the bunch of curving timbers that formed so conspicuous a feature of the stern of the classical trireme.

²¹ ‘Anon PBPP,’ 2, 6.

²² Cameniata, op. cit., c. 32. The same qualities may be inferred from a passage in Comnena, Anna, Alexiad, XI, 10, 4Google Scholar, where the impact of a ramming Dromon was not sufficient to snap the oar or its lashings.

²³ When constructing my model I experimented with probable positions for this gangway, and found that the arrangement described presented the fewest difficulties.

²⁴ ‘Anon PBPP,’ 2, 8.

²⁵ Alexiad, VI, V. 7. A wale so placed would help to explain the peculiar position of the ballast.

²⁶ There had been an alteration in the name, and they were now known as μανικέλλια, ‘Anon PBPP’, 2, 12.

²⁷ It is clear from the ‘Anon PBPP’ that there was no form of apostis.

²⁸ In a saline solution iron and lead set up mutual corrosion. This simple fact was appreciated by the Romans but later forgotten. It was rediscovered in 1682; cf. Chatterton, E. Keble, Sailing Ships and their Story, 77–9Google Scholar.

²⁹ cf. Prod., Theod., Rhodanthe et Dosicles, V, 444Google Scholar; Comnena, Anna, Alexiad, VI, 5, 7Google Scholar.

³⁰ The ordinary Mediterranean anchor was only too often a large stone with a hole in the middle.

³¹ Περὶ θαλ. 6; ‘Anon PBPP,’ 2, 10.

³² It is surprising how light a missile will penetrate the bottom of a boat. In the eighteenth century privateers would beat off boarding parties from men-of-war by sinking their boats with hand-thrown shot weighing no more than nine or so pounds.

³³ It seems wrong to appeal to the authority of De caerimonüs 672, for this passage is an inventory of material supplied by only one government department. The vital passage in Leo is Περὶ θαλ. 7. I owe the second of the suggested emendations to Professor Jenkins, who points out that it does less violence to the text. Alternatively, we might suppose that μέσον represents the Arabic ‘mizan’, the source of the English ‘mizzen’. The Arabic form might well have been transliterated μέσον, and the emendation to μέσον would have appealed to any intelligent copyist. My main argument, however, would not be affected, and it may be noted that the after-mast of a two-masted galley was stepped approximately amidships: cf. the plates in Furttenbach's Architectura Navalis of 1629.

³⁴ Alexiad, X, 8, 1.

³⁵ C. 32. The whole chapter is of the greatest importance, and with some of its implications I hope to deal elsewhere.

³⁶ Torr, op. cit., 91, n. 197.

³⁷ A dictum of Pantero Pantera is very apposite: ‘Portano le galee ordinariamente doi arbori, il maestro et il trinchetto.’ ‘Galleys have usually two masts, etc.,’ Armata Navale (Rome, 1614), 46Google Scholar.

³⁸ ‘Anon PBPP’, 2, 10.

³⁹ The advantages of composite construction can best be studied in the plans of fishing boats printed in Paris, F. E., Souvenirs de Marine (Paris, 1882–1886)Google Scholar.

⁴⁰ Nic. Uranus, c. 121 (ed. Dain).

⁴¹ Leo, Περὶ θαλ. 41–8.

⁴² For the use of cones, see Leo's mention of καμελαύκια. The fact, however, that they were hoisted on a spear suggests that this particular signal was obsolete. If still used, καμελαύκια would have been hoisted in the rigging.

For heliographs, see Syrianus Magister, 7, 1 (ed. Dain).

⁴³ During the tenth century there was a tendency to increase the number of flame-throwers, and in 949 Dromon's carried three. Some form of propellent explosive seems to have been used, perhaps crude gunpowder.

⁴⁴ Leo, Περὶ θαλ. 60.

⁴⁵ Professor Adcock brings to my notice an interesting anticipation in Justin XXXII, 4, 6—was Leo familiar with a Greek translation or with Justin's lost source ?

⁴⁶ Leo, Περὶ θαλ. 71.

⁴⁷ Ibid. 82. The Dromonia of Constantine Porphyrogenitus, however, are full-size Dromons. His misuse of the diminutive can only reflect demotic usage.