114 A NATURALIST ON DESERT ISLANDS. 



nature across an open sea could only have been rendered 

 possible by the rats having been carried out to sea by 

 means of a floating island of vegetation, or a mass of 

 entangled tree-trunks such as one often sees in mangrove 

 swamps. 



The first Swan Island rat we saw was in the hands of Mr. 

 Eagle, the first officer, who had just shot it. As he came 

 up, holding it by its short tail, it looked for all the world 

 like a half-grown rabbit. The second ran over some rocks 

 in front of us, and tried to escape under the spreading 

 roots of a large undermined tree. This, after a certain 

 amount of caution, for we did not then know that these 

 formidable-looking rats were surprisingly mild mannered, 

 I caught by a hind leg, and drawing it carefully out, 

 gradually enveloped it in my felt hat. From this it was 

 transferred to a fishing creel, and we went on our way 

 rejoicing. 



These rats are diurnal in their habits ; and we saw at 

 least a dozen others running about and bolting into the big 

 crevasses with which the island is seamed. We took 

 several other specimens * and captured one more alive. 

 One of these live specimens eventually arrived safely 

 in England, and was in all probability the first example 

 of the genus to do so. 



I had the honour of exhibiting it one day, not only to 

 his late Majesty, King Edward VII., but to his favourite 

 dog " Caesar." The rat exhibited not the faintest signs 

 of fear or suspicion in the presence of the dog, or even of 

 awe in the presence of His Majesty ; and it seemed inter- 

 esting to reflect, that if we exclude ourselves, " Caesar " 

 was the very first mammal, of any kind other than his own 

 race, that the rat had ever set eyes on. While the dog 

 was carefully held back, the rat came to the edge of the 

 low table it was on, and patting its muzzle within an 

 inch or two of the dog's, quietly inspected, what to it 



* One of these ra3asured 13J inches from, tip of snout to base of tail, 

 and 16| inches to end of tail. Owing to its niassive hind-quarters and 

 thick-set head the animal appears much larger than these measurements 

 would suggest. 



