I read an interesting article earlier today (found here) on how our probable conclusions for triremes have been wrong. I found the article firstly very enlightening and secondly very interesting as well: there’s the question of what a triere actually looked like, but this coexists with our wish to impose our own thoughts on the ancient concept.
So, the suggestion goes: men were three abreast in total, outside rows rowing either side and the centre row rowing both sides. It is an interesting idea, and sounds like a reasonable compromise between available manpower and the strength of a single ship (although how can we guess what Athenians thought reasonable?).
Now, one has to note that the idea of the oars all being on one level (or, sometimes on two) with men in three rows is one of relative simplicity, especially when considering the alternatives that are considered in the same topic (three separate levels of oars). What follows from simplicity is that it would also have been the easiest way to train the men to be in rhythm, and without rowing in rhythm much of the energy would have been wasted. Simplicity really has to be the key for things to succeed in building a fine ship.
And that is one of the reasons for which I quite like this new theory.