ANTS AND SOME OTHER INSECTS. 45 



Can we generalise to such an extent as to apply this term with- 

 out further investigation to all arthropods? To a considerable ex- 

 tent this must be denied. 



In fact, the multiformity in the structure and development of 

 the arthropod sense-organs is enormous, and we must exercise cau- 

 tion in making premature generalisations. 



It is certain that in some aerial insects the olfactory sense has 

 dwindled to a minimum, e. g., in those species in which the male 

 recognises and follows the female exclusively by means of the eyes, 

 as in the Odonata (dragon-flies). To insects with such habits an 

 olfactory sense would be almost superfluous. Here, too, the an- 

 tennae have dwindled to diminutive dimensions. 



But there are insects whose antennas are immovable and quite 

 unable to touch objects. This is the case in most Diptera (flies). 

 Still these antennas are often highly developed and present striking 

 dilatations densely beset with olfactory papillae. By experiment I 

 have demonstrated the existence of an olfactory sense in such 

 Dipteran antennae, and I have been able to show that, e. g., in 

 Sarcophaga vivipara and other carrion flies, the egg-laying instinct 

 is absolutely dependent on the sensation of the odor of carrion and 

 the presence of the antennae. In these cases the contact-odor sense 

 is undoubtedly absent. More or less of a topochemical odor-sense 

 at long range must, of course, be present, since the antennae are 

 external, but the precision of the spatial image must be very im- 

 perfect, owing to the immobility of the antennas. Nevertheless, 

 flies move about so rapidly in the air that they must be able by 

 means of their antennae to distinguish very quickly the direction 

 from which odors are being wafted. These insects do, in fact, find 

 the concealed source of odors with great assurance. But this is no 

 great art, for even we ourselves are able to do the same by sniffing 

 or going to and fro. But the flies find their way through the air 

 with their eyes and not at all by means of their sense of smell. 

 Hence their olfactory powers probably constitute a closer psycho- 

 logical approximation to those of mammals than to the topochemi- 

 cal odor-sense of ants, for they can hardly furnish any constant and 

 definite space-relations. 



