BREEDING OF DRAGON FLIES. 89 



lived insects falling behind furniture or other inaccessible 

 places would attract a brood of Derniestes, the next gener- 

 ation of which would be nourished on the family furs. 



Before drawing any conclusions upon the foregoing, it 

 is necessary to reply to some of the premises set forth in 

 the circular. The question is raised, wliy it is not as easy 

 to breed dragon flies as it is fish, silk worms, bees, etc.? 

 Let us ask ourselves the question, why it is not as easy to 

 produce herds of tigers, lions, and some other carnivorous 

 and herbivorous quadrupeds as it is cattle, sheep, and 

 horses? We know that there is a diiference; that this 

 difference is inherent in the habits and instincts of these 

 animals as shown by experience; and yet why this differ- 

 ence no one will ever be able to satisfactorily explain until 

 tlie relation between the organs of generation and the in- 

 stinctive action of the motor nerves can be discovered. 

 As among quadrupeds there are species which can be 

 easily bred in captivity from generation to generation, and 

 others which cannot, so there exists among insects the 

 same difference. To illustrate : Of the nocturnal Lepidoj)- 

 tera, nearly all the species of Bombycidae readily mate in 

 captivity. So much so, that in the case of the silk worm 

 (Bombyx mori) pairing will inevitably occiu' if specimens 

 of both sexes be confined in any space, however limited, 

 oviposition will as surely follow upon any material within 

 a definite period thereafter, irrespective of fertilization. 

 The same fact is true of a large number of the silk spin- 

 ners, such as Callosamia promethea, Hyperchiria io, and 

 Samia cynthia (Japanese silk moth, feeding upon the Ail- 

 anthus, acclimated here), while on the other hand the 

 species of Sphingidae and many of the Xoctuidre and 

 Geometridfe can only be mated with extreme difficulty. 



