216 RAMBLES ABOUT HOME. 



sional failures to seize their prey. This seemed to con- 

 firm my original impression, but, as a late ornithologist 

 of note has said, " The horizon of one man is at best 

 very limited, and many ornithological facts occur that are 

 not dreamed of in his philosophy." I repeated my observa- 

 tions through the spring and summer of the ensuing year. 

 My opportunities were equally good, and, much to my 

 satisfaction, I have a quite different story to relate. It 

 is proper, however, to state that during the summer of 

 1873 my observations were made altogether in a very 

 limited locality the summit level of a canal and were 

 confined to one pair of birds. During the subsequent 

 year, I watched the kingfishers in several different locali- 

 ties, and my note-books make mention of these birds 

 from two to six times per day, for one hundred and one 

 days, or a total of about four hundred observations. Of 

 this series I have to say that in eighty-eight instances the 

 kingfisher captured and, alighting, deliberately beat the 

 fish against the limb of the tree, and afterward swallowed 

 it. Thus it will be seen that this habit is by no means 

 constant, as less than one fourth of the fish taken were 

 killed before being swallowed ; though, on the other hand, 

 it is evident that I was wide of the mark in stating that 

 the fish is always swallowed without being first killed. 



There is, of course, some cause for this difference in 

 the habits of these birds, and I believe it may be ex- 

 plained in this way : as already stated, my observations 

 during 1873 were confined to a single pair of these birds, 

 in one locality ; and the obvious reason why these par- 

 ticular kingfishers always swallowed their prey as soon as 

 caught was because they fed exclusively on the very small 

 but extraordinarily numerous cyprinoids frequenting this 

 artificial sheet of water. I know, of my own fishing ex- 

 perience (pursued after a different manner, however), 



