The Byzantine admiral, encouraged by the arrival of his allies, then sallied forth from the Heptascalon, and led the way towards the Genoese ships at Chalcedon. The latter, finding it impossible to make head against the wind, retired towards Galata, and skilfully entrenched themselves among the shoals and rocks off Beshiktash, preferring to be attacked in that advantageous situation.[1152] The allies came on, and a desperate conflict, partly on the water, partly on the rocks, ensued, until night parted the combatants without a decisive victory on either side. With this narrative of Cantacuzene in view, no one familiar with the vicinity of Constantinople can doubt for a moment that the Neorion at the Heptascalon was upon the Sea of Marmora. The single circumstance that the walls in the neighbourhood of the harbour were protected by boulders placed in the sea as a breakwater is alone sufficient to prove the fact; for only the walls bordering the Sea of Marmora were defended in that manner. Equally conclusive is the circumstance that the Venetian and Spanish ships found it difficult to make the harbour from the Prince’s Island with a strong south wind on their left. Such a wind would drive them towards the Bosporus with a violence that would render it almost impossible for them to put into any port on the Marmora shore of the city. Nor is it less decisive to find, as the historian’s account makes perfectly clear, that the harbour was so situated; that the approach to it, and possible shipwrecks at its entrance, could be observed by the Genoese admiral stationed off Chalcedon; that an enemy at Chalcedon found it hard to advance towards the Heptascalon in a strong south wind; and that vessels proceeding from the harbour to Galata could, on the way, touch at Chalcedon. These facts hold true only of a harbour on the Sea of Marmora.