Borimir Jordan and Alec Tilley

Dr Morrison recently published his opinion (Shaw 1993: 19) that neither the Olympias nor any other ship with six banks of oars could conform with the evidence on ancient triremes. In that we agree. For a time it seemed that the long-running trireme controversy might be concluded except for the final and probably irreconcilable difference between those who, like Dr Morrison, reject the evidence (of Galen and 'Aristotle') in favour of their hypothesis, and those who do the opposite.

Dr Coates (1995) maintains, despite the evidence of the Olympias, that ancient triremes had six banks of oars. He does not dispute the fact that the Olympias fails to conform with the many items of ancient evidence cited in Tilley (1992). He asserts (p 161) only that the Olympias conforms to 'the more generally accepted ancient evidence' but does not say what that is. We do not think that, true or false, his assertion is of any value: in six-banked-trireme literature, evidence which does not conform with the hypothesis is never accepted. It is simply ignored, rejected or 'improved' until it does conform.

Nor does Coates deny that the arrangements of oars put forward in Tilley's article conform exactly with the ancient written and iconographic evidence, claiming only that they fail to accord with 'generally accepted interpretations of ancient evidence'. An examination of Figure 1 will show the important distinction between interpretation and evidence, and the way in which Coates 'improves' ancient evidence that does not accord with his notions.

Dr Coates's interpretation of the fragment in Figure 1 (perhaps one of the 'generally accepted interpretations' he alludes to) is of oars arranged as in the Olympias. The Olympias's uppermost level of oars is suggested to him by a line of thole pins which through 'rough drawing' (Morrison and Coates 1986: 150) the artist has entirely neglected to depict. He interprets the lower-level oarport as a large, circular, Olympias-type oarport, transmuted by the same 'rough drawing' into the equal-sized, semi-circular oarport that we can actually see. Thus the interpretation is amazingly like the Olympias. If on the other hand one works directly from the evidence, ignoring interpretation, one can see three equal-sized, semi-circular oarports, arranged in a way that is suitable for rowing triple-banked (Tilley 1992 FIG. 6, 7 and 8). Arguing from evidence and arguing from interpretation are very different things, as different as astronomy from astrology.

Other ancient representations which do not show oars at three levels have received equally fanciful treatment and presented as pseudo-evidence for six-banked triremes, notably the Lenormant relief (Morrison and Coates 1986: 139-41).

We consider that the Olympias rows badly enough to suggest that she is unlike a trireme in respect of performance under oars. Her best published speed over a mile is 7.05 knots with a flying start, appreciably slower than the 7.8 knots that an ancient trireme could, according to Morrison and Coates (1986: 103), row for sixteen and a half hours to cover 129 miles in a day, with no indication of haste. Coates's statement that the Olympias 'has touched 8.9 knots, which represents a shortfall of about a knot from the ancient sprint speed likely to have been sustainable for five minutes, requiring an increase of about 35% in power to rectify' (p 161) could not lawfully be included in a company prospectus. The juxtaposition of an instantaneous speed and an unstated speed maintained for five minutes is misleading and the figure 35% unreal, because the Olympias has not sustained 8.9 knots for five minutes.

In support of the theory that ancient triremes had six banks of oars, Coates cites the six-banked Venetian galleys called triremi, though he knows they were given that name by classicists who thought ancient triremes had six banks of oars (Morrison 1941: 18-19). He suggests (p 160) that to describe a double-banked boat after the number 'two' is peculiar to Tilley or (p161) to northern Europe, though Mediterranean seamen have used the terms armeacute; en couple, doppio ordine di reme and diplokopos to describe a double-banked boat for generations. We are grateful to Professor Maurice Pope for a citation that, together with the linguistic evidence not challenged by Dr Coates, should settle the question of ancient rowing nomenclature. The Latin poet Manilius wrote about a man swimming:

'nunc alterna ferens in lentos bracchia tractus ... nunc aequore mersae/ Diducet palmas furtiva biremis in ipso' (MAN. 5, 424 - 425)
which Goold translates:
'Now lifting one arm after the other to make slow sweeps ... now like a hidden bireme he will draw apart his arms beneath the water'.

That second swimming action is the breast stroke. A man's arms swimming breast stroke are like the oars of a double-banked boat. They give no suggestion of Dr Coates's idea of a bireme, a vessel with oars at two levels.

We agree with Coates that his assertion of the impracticability of three-banked trireme reconstructions ought to be put to the test, using one or both of the systems proposed in Tilley (1992) rather than those ascribed to him in Coates (1995). We suggest the following sequence of trials:

  1. a. A full-sized mock-up of the bows of the oared ship that supports the Nike of Samothrace. This would, we believe, show that it is suitable for triple-banked rowing as shown on the Siren Vase, and not for any other method.
  2. b. More thorough triple-banked rowing trials, which would, we believe, demolish Coates's objections.
  3. c. 'Reconstruction' of a double-banked pentekontor; first on paper, then as a tank-test model, before a full-sized ship was attempted.
  4. d. Conversion of the pentekontor to row triple-banked in the Siren Vase manner. We believe the conversion would give a useful increase in speed.

Support for those experiments from the Trireme Trust would be advantageous, if its members were willing to extend their area of interest to include triple-banked rowing.


  1. COATES, J.F. 1995. Tilley's and Morrison's triremes - evidence and practicality, Antiquity 69: 159-62.
  2. MANILIUS, Astronomica, ed. G.P. Goold, Loeb Classical Library (Cambridge, Mass., London 1977).
  3. MORRISON, J.S. 1941. The Greek trireme, Mariner's Mirror 27: 14-44.
  4. MORRISON, J.S. and J.F. COATES 1986. The Athenian trireme. Cambridge: Cambridge University Press.
  5. SHAW, J.T. (ed) 1993. The trireme project. Oxford: Oxbow Books.
  6. TILLEY, A.F. 1992. Three men to a room - a completely different trireme, Antiquity 66: 599-610.

Fragment of red-figure pottery dated c. 450 BC.

Figure 1 - Fragment of red-figure pottery dated c. 450 BC.