A genuine topochemical antennal sense is, therefore, probably best developed in all arthropods, whose antennæ are not only movable in the atmosphere, but adapted to feeling of objects. In these cases the still imperfect topochemical odor-sense for distances can be momentarily controlled by the contact-odor-sense and definitively fixed topographically, i. e., topochemically, as we see so extensively practised in the ants. It would be possible to meet this view with the objection that a contact-odor sense could not accomplish much more than the tactile sense. I have made this objection to myself. But in the first place it is necessary to reckon with the facts. Now it is a fact that insects in touching objects with their antennæ mainly perceive and distinguish the chemical constitution of the objects touched and heed these very much more than they do the mechanical impacts also perceived at the same time. Secondly, the tactile sense gives only resistance and through this, form. On the other hand, the multiplicity of odors is enormous, and it is possible to demonstrate, as I have done for the ants, and Von Buttel-Reepen for the bees, that these animals in distinguishing their different nest-mates and their enemies, betray nothing beyond the perception of extremely delicate and numerous gradations in the qualities of odors. In combination with topochemical space-perception, these numerous odor-qualities must constitute a spatial sense which is vastly superior to the tactile sense. The whole biology of the social Hymenoptera furnishes the objective proof of this assertion. It would certainly be well worth while to investigate this matter in other groups of arthropods which possess complex instincts. In conclusion I will cite an example, which I have myself observed, for the purpose of illustrating the capacity of the topochemical olfactory sense. The American genus Eciton comprises predatory ants that build temporary nests from which they undertake expeditions for the purpose of preying on all kinds of insects. The Ecitons follow one another in files, like geese, and are very quick to detect new hunting grounds. As “ants of visitation,” like the Africo-Indian species of Dorylus, they often take possession of human dwellings, ferret about in all the crevices of the walls and rooms for spiders, roaches, mice, and even rats, attack and tear to pieces all such vermin in the course of a few hours and then carry the booty home. They can convert a mouse into a clean skeleton. They also attack other ants and plunder their nests. Now all the workers of the African species of Dorylus and of many of the species of Eciton are totally blind, so that they must orient themselves exclusively by means of their antennal sense.