STATE HORTICULTURAL SOCIETY. 41 



The principal difficult)' of the investigation appeared when I began 

 to estimate the amount of injury or benefit indicated by the contents of 

 the stomach of each bird. Before I could tell how much credit to give 

 to a bird for eating beetles indiscriminately, I must know what ratio the 

 injurious beetles held to the beneficial and the neutral ; and when I found 

 the food composed of a dozen different elements — caterpillars, beetles, 

 hymenoptera, blackberry seeds, etc. — I must settle in my own mind what 

 relative importance to give to each of these elements. 



Of course, all this must be a matter of individual knowledge and 

 iudgment. The problem is not one which can ever be settled with entire 

 accuracy. All we can expect is, that our conclusions shall approximate 

 the truth more and more closely as observations are multiplied. Investi- 

 gations of this kind are well illustrated by the prognostications of the 

 Weather Bureau. If observations are carefully made and correctly tabu- 

 lated, predictions based upon them increase in certainty with the number 

 of the data. 



In making up my estimates of the horticultural value of the different 

 species of birds, I based my calculations upon the following guiding i)rin- 

 ciples, to which I ask your attention, because the value of the estimates 

 depends very largely upon the correctness of these principles : 



1. Any bird, of which it is only known that it feeds upon insects, is to 

 be regarded as beneficial, until facts are discovered to the contrary. (1 

 have heard it argued that a bird feeding indiscriminately upon insects of 

 all kinds might do more harm, by destroying insectivorous and parasitic 

 forms, than good by devouring the injurious ones. A moment's reflection 

 will show the unreasonableness of this view. The extermination of all 

 insects, of every species, would doubtless be a benefit to horticulture, 

 although by no means unmixed with 6vi\. The destruction of half the 

 individuals of each species would therefore be proportionably beneficial, 

 and the same must be said of a hundredth, a thousandth or a millionth 

 part — or, in short, of any proportion of them eaten by birds having no 

 preferences. As the beneficial insects are, as a rule, either smaller or 

 more active (or both) than their victims, they will be, in fact, less liable 

 to capture by birds than the latter. 



2. A bird feeding upon hymenoptera is to that extent probably inju- 

 rious. If the hymenoptera found are neither Tenthudinidce nor Neveri- 

 d£e, this probability is greatly strengthened. If they are known to be 

 Ichneumonidre, it becomes a certainty. A bird feeding on ants is to that 

 extent neutral. 



3. A bird feeding on lepidoptera is to that extent probably benefi- 

 cial. If this is a twilight bird, it is almost certainly .so. 



4. A bird may be reckoned beneficial in so far as it feeds upon cater- 

 pillars with two rows of abdominal prolegs. 



5. I can infer little or nothing at present from the presence of diptera 

 in a bird's stomach. 



