_Q._—Assuming there was a grating over the stoke-hole unprotected, would that be a cause of danger? _A._—Unless properly protected, it would be a serious cause of injury. It is the ordinary custom for all hatches to be protected by tarpaulins, cleated and battened down. It would be a reasonable and proper thing to make a provision for tarpaulins over that grating, and it should be compulsory to have such fastenings. I have not seen the window in the fore part of the poop, but if there was, it should have had a shutter outside on hinges capable of being fastened inside and out. It is customary on all new ships to have the bunker hole covers secured, and so they ought to be. Cross-examined by Mr. Mitchell: I have been on the _Alert_ I daresay one hundred times doing repairs to the ship. It was an improper thing not to be able to attach a cover to the grating. It would be a very heavy sea that would go over it from the weather side, but if the vessel were thrown on her beam ends, there is nothing to prevent it going in on the lee side. An angle of 45 degrees would be sufficient to throw her over enough to take in a lee sea that would cover it. It is not the throw over of the vessel alone, it is simultaneously the rise of the sea to leeward. I have no knowledge how far the ship did go over on this occasion, but I think there would not be much of a sea from the weather side. If she did take water in on the weather side, it would be principally forward of the bridge, and it would get to the lee side before getting to the grating. By 45 degrees I mean the vessel has to go 45 degrees from the perpendicular. This is not an unusual angle for a ship to heel over; in my experience I have known them to heel over almost to a right angle. His Honour,—I have known large steamers to roll over to a greater degree than an angle of 45. _Q._—By Mr. Mitchell: What is your idea as to how the _Alert_ was lost? _A._—The opinion I have formed is that they had a change of wind and a mixed sea; that the vessel began to labor and take in much water, which eventually took her down. On measuring the model of the _Alert_ now shown in court, I find that the vessel’s length was eighteen times her depth and nine times her beam, approximately.