The afternoon saw the Grampus well across the gulf, and by five o'clock she changed her course to south by east, leaving the densely wooded hills of Trinidad far behind with the coast of Venezuela in plain view to starboard.

Ah Sin, having been duly instructed as to his duties, prepared the supper on the electric stove, and served it. Speake relieved Matt at the steering gear, and when Dick went below to take Gaines' place at the motor, Matt sprawled out on the locker to catch his own forty winks. A stiff sea was running, and the Grampus was submerged to a depth that merely left the periscope ball clear of the combers.

As the darkness deepened, Speake had Carl put the turbines at work, throwing out sufficient water ballast to lift the conning-tower lunettes clear of the waves. The electric projector was then turned on, and a ray of light shot through the forward lunette and marked the submarine's path through the tumbling sea.

For some hours everything went well. Then abruptly the motor began to sputter and misfire, lessening the speed of the boat and throwing her—now that she was riding higher and with the top of the conning tower awash—more at the mercy of the waves.

Loose furniture began to slam around the periscope room. Matt was thrown from the locker, and sat up, wondering what had gone wrong with the motor.

"What's the matter down there, Dick?" he called through the motor-room tube.

"I'm a Feejee if I know," Dick answered. "You'd better come down, old ship, and take a look."

Matt was soon at his chum's side. His keenly trained ear was usually able to locate any ordinary trouble, but this time he was puzzled. The ignition was all right, and the supply pipe from the tank was clear. Nevertheless the motor sputtered and jabbered with a wheezy but unsuccessful attempt to do its full duty. The platinum, in the blade or spring of the commutator, will, in rare cases, get loose and cause misfiring, but that was not the cause of the present trouble. Another rare cause, resulting in similar symptoms, lay in the loosening of the carbon pole in the cell of a battery. But, just now, the batteries were not at fault.

Finally, as a last resort, Matt examined the gasolene that was being fed into the carburetor. A few drops in the palm of his hand aroused his suspicions. The next moment the hydrometer test was made and water was found in the gasolene.

"How did it get there?" demanded Dick. "The gasolene has worked well enough all afternoon and so far during the night."