298 PROCEEDINGS OF THE ACADEMY OF [Apr.. 



of a large worker of Formica or of Camponotus. This caused the 

 ant to discharge all of the formic acid stored up and some of it lodged 

 on the tip end of the abdomen. The liquid on the abdomen gave off 

 a penetrating odor which lasted four or five minutes. 



An ant or hornet was carefully removed from its nest or cage and 

 was placed into one of the experimental cases. When first put into 

 the case the insect usually wandered about for several minutes, but 

 finally became quiet. The insect was tested with each of the above 

 odors only wdien it became perfectly quiet, without even the antennae 

 being moved in the least. The stopper of a vial was quickly removed 

 and the vial was gently and slowly placed under the experimental 

 case directly beneath and within one-half inch of the individual 

 being tested. Occasionally the vial was placed a few inches in front 

 of the specimen, to test its ability of smelling for a short distance. 

 When using the odor from the formic acid, the ant whose abdomen 

 carried some of this acid was held by a pair of forceps under the 

 experimental case in the same position in which a vial was held. 

 When all of these precautions were taken, a normal ant or hornet 

 responds to anyone of these odors without failure. As a control, 

 an empty and odorless vial was now and then placed under the 

 insects in the same manner. If by chance an ant or hornet moved 

 while the control test was being made, its behavior was quite different 

 from that observed when odors were used. Only the first responses 

 have been recorded and in all cases where there was the least doubt 

 as to whether the insect moved for any reason other than the olfactory 

 stimulus, such movements were never recorded. The reaction time 

 was counted in seconds. With an ordinary watch the minimum time 

 which can be definitely recorded is two seconds, although most of 

 the individuals responded to some of the odors much more promptly. 

 Owing to this source of error, the average recorded time is probably 

 double what it should be in the cases when the response was prompt. 

 An intermission of 10 minutes elapsed between any two tests in the 

 same experimental case. A few individuals were tested twice with 

 the same odor, l)ut most of them were tested only once. 



In the following paragraphs are given the responses of the three 

 castes of ants and of work;>r hornets to the odors of the six different 

 substances and the average reaction times in seconds. In recording 

 the responses the term "vibrated" is used to describe the rapid 

 movement of the antennae up and down or from side to side. When 

 this movement is slow they are described simply as having ''moved. " 

 Quite often an ant or hornet lies flat on its thorax and abdomen, 



