126 PASSERES. TROCHILID^. 



deavoured to replace them, by sending to the moun- 

 tain ; for where the species was so numerous two 

 months before, they were now (beginning of June) 

 scarcely to be seen at all. The cause of the death 

 of my caged captives, I conjecture to have been 

 the want of insect food ; that, notwithstanding their 

 frequent sipping at the syrup, they were really 

 starved to death. I was led to this conclusion, by 

 having found, on dissecting those which died, that 

 they were excessively meagre in flesh, and that 

 the stomach, which ordinarily is as large as a pea, 

 and distended with insects, was, in these, shrunken 

 to a minute collapsed membrane, with difficulty 

 distinguished. If I had an opportunity of trying 

 the experiment again, vdth the advantage of this 

 experience, I would proceed rather differently. 

 I would have a very capacious cage, wired on every 

 side, in the bottom of which a supply of decaying 

 fruit, such as oranges or pines, should be constantly 

 kept, but covered with wire that the birds might 

 not defile their plumage. This, as I have proved, 

 would attract immense numbers of minute flies, 

 which, flitting to and fro in the cage, would pro- 

 bably afford sufficient sustenance to the birds in 

 conjunction with the syrup. The birds, however, 

 should be caged as short a time as possible before 

 sailing, which might be early in May ; and by a 

 steamer, which calling at St. Thomas, Bermuda, and 

 the Azores, large bunches of fresh flowers, and even 

 herbage, might be obtained at short intervals on 

 the voyage, with which, of course, a multitude of 

 insects would be introduced. Thus, I still think, 



