[122] Small informed Watt that it was intended to make an engine for the purpose of _drawing canal boats_. “What Mr. Boulton and I,” he wrote, “are very desirous of is, to move canal boats by this engine; so we have made this model of a size sufficient for that purpose. We propose first to operate without any condenser, because coals are here exceedingly cheap, and because you can, more commodiously than we, make experiments on condensers, having several already by you. Above 150 boats are now employed on these new waveless canals, so if we can succeed, the field is not narrow.” This suggestion of working canal boats by steam immediately elicited a reply from Watt on the subject. Invention was so habitual to him that a new method of employing power was no sooner hinted than his active mind at once set to work to solve the problem. “Have you ever,” he wrote Small, “considered _a spiral oar_ for that purpose, or are you for two wheels?” And to make his meaning clear, he sketched out a rough but graphic outline of a screw propeller. Small’s reply was unfavourable: he replied, “I have tried models of spiral oars, and have found them all inferior to oars of either of the other forms; I believe because a cylinder of water immersed in water can be easily turned round its own axis. We propose to try gun-lock springs with the fixed part longer than the moving. If we cannot succeed, we will have recourse to what you have so obligingly and clearly described.” Finally Watt writes a fortnight later, “concerning spirals, I do not continue fond of them.” About the beginning of 1770, another unsuccessful experiment was made by Watt and Roebuck with the engine at Kinneil. The cylinder had been repaired and made true by beating, but as the metal of which it was made was soft, it was feared that the working of the piston might throw it out of form. To prevent this, two firm parallel planes were fixed, through which the piston worked, in order to prevent its vibration. “If this should fail,” Roebuck wrote to Boulton, in giving an account of the intended trial, “then the cylinder must be made of cast-iron. But I have great confidence that the present engine will work completely, and by this day se’nnight you may expect to hear the result of our experiments.”[123] The good news, however, never went to Birmingham; on the contrary, the trial proved a failure. There was some more tinkering at the engine, but it would not work satisfactorily; and Watt went back to Glasgow with a heavy heart. [123] Roebuck to Boulton, February 12, 1770.