2 THE BERMUDA ISLANDS. 



was visible. As we approached nearer to the shore, however, 

 abrupt changes in the color of the water revealed the position 

 of the coral-shallows, but we as yet saw nothing of the reef of 

 the imagination. Occasional jelly-fishes floated lazily by, and 

 the ever merry petrels were still quivering in our path of 

 beaten foam. Our first tropic-bird hovered about the rigging, 

 seemingly surprised at the early intrusion which we had per- 

 mitted ourselves. The flying-fishes became more numerous as 

 we neared the islands, and they could be frequently seen skip- 

 ping away five or more at a time, and usually in a direction at 

 right angles to the line of the ship. 



I was anxious to determine the true nature of their aerial 

 locomotion, and 'to settle the vexed question of the supposed 

 flying movements of the fins. We accordingly watched these 

 interesting creatures very intently, and followed them with our 

 glasses over their entire course. So rapid was their motion, 

 however, that it was not easy to keep them within the field 

 of the glass, and still less easy to hold them in distinct vision, 

 and for a long time we really hardly knew what we saw. 

 We failed, however, to detect any positive continuous move- 

 ment on the part of the fins, and it certainly appeared as 

 though in all, or nearly all, cases the animal merely shot 

 forward as the result of some primary impetus, taking a 

 course nearly horizontal with the surface of the water. 

 This direct course, in view of the apparent method of pro- 

 pulsion, was certainly surprising, as it would naturally be 

 expected that from an initial leap the line of travel would be 

 that of a regular curve. But the horizontal course was pos- 

 sibly more apparent than real. At intervals, two or three times 

 in a single flight, the animal seemed to strike the crest of a 

 wave, and acquire new impetus from a blow of the tail. In 

 this way the flight may be said to be one of distinct passages 

 although without arrest of movement in which the curve 

 motion largely disappears, or is at least reduced to its lowest 

 terms. Indeed, the successional character of the flight could be 

 plainly seen in the sudden jerky changes of direction which 



