The Food of Birds. 109 



cial test of the conclusion implied above. In the paper 

 quoted from, I gave the details and a summary of the food 

 of forty-one robins in a table similar to those presented in 

 this paper, and a comparison of the averages of that table 

 with those of the table on pages 112, 113, 114, 115, may be 

 easily made. While any serious differences in the averag- 

 es of these two tables would not necessarily condemn the 

 later one, but, at the worst, would leave its sufficiency in 

 doubt, a substantial agreement of the two would be con- 

 clusive proof of the correctness of both. It is incredible 

 that the averages of a hundred and fourteen specimens 

 should agree essentially with those of forty-one, unless 

 both were framed upon identical principles and were suffi- 

 ciently true to the facts for all practical purposes. I will, 

 therefore, place the principal averages of these tables side 

 by side, premising that the later table not only includes 

 nearly three times as many specimens as the earlier, but 

 covers two months' more time. 



The figures for the first and second tables, taken alter- 

 nately, are as follows : — Insects, seventy per cent, and six- 

 ty-five per cent. ; caterpillars, eighteen per cent, and sev- 

 enteen per cent. ; Diptera, eighteen per cent, and seven- 

 teen per cent. ; Ooleoptera, nineteen per cent, and eight- 

 een per cent. ; Carabida?, seven per cent, and five per 

 cent. ; Scarabteida?, four per cent, and seven per cent. ; 

 Lachnosterna, two per cent, and three per cent. ; Elateri- 

 dae, three per cent, and two per cent. ; Rhyncophora, three 

 per cent, and two per cent. ; Chrysomelidse, one per cent. 

 and a trace ; Hemiptera, four per cent, and three per cent. ; 

 Orthoptera, eight per cent, and four per cent. ; Arachni- 

 da, a trace and one per cent. ; Myriapoda, two per cent, 

 and a trace ; garden fruits, twenty-eight per cent, and 

 twenty-nine per cent. 



As I did not discriminate, in the former table, between 

 tame and wild edible fruits, I have included the latter in 

 both, and excluded the inedible fruits. I believe that the 

 agreement in these figures, taken into account the earlier 

 and later months covered by the second table, is quite re- 

 markable, and can be explained only on the supposition 

 that the fuller table presents a reasonably accurate sum- 



