74 
Psyche 
[December 
able from that caused by direct mechanical disturbance. Other body 
parts tested in this manner do not induce the same reaction. Al- 
though the volatile components of the anal glands of each species pro- 
duce a distinctly different smell to humans, there is evidently no 
species-specificity shown in the alarm response. Unlike the complete 
specificity shown in responses to trail substances, squashes of anal 
glands of each species produce approximately the same intensity of 
alarm curiosity in workers of the other two species as in its own. 
These results can be interpreted as indicating that either a common 
component exists in each of the anal secretions or else, as seems more 
likely, the secretions differ but the reactions are not species-specific. 
In contrast, disturbed workers of Iridomyrmex humilis do not re- 
lease quantities of volatile substances sufficient to be detected by the 
human observer. Moreover, they do not respond with more than mild 
excitement to single anal-gland squashes from sister workers or the 
other three dolichoderine species. 
Pavan and Trave (1958) have shown that the volatile components 
of anal gland secretion of Tapinoma nigerrimum (Nylander) are 
methylheptenone (C 8 H 14 0 ) and propyl-isobutyl-ketone (C 8 H ie O). 
Since our initial experiments suggested that the action of alarm sub- 
stances is not species-specific, it was not surprising to find that small 
quantities of the two nigerrimum substances, manufactured synthet- 
ically and presented in separate tests, caused intense alarm behavior 
in Tapinoma sessile. No attempt has been made to map the full range 
of compounds that will induce the alarm frenzy in the dolichoderines 
that communicate alarm by anal secretions. However, the following 
preliminary study will show that not all volatile ant secretions have 
this effect. A simple olfactometer test designed to measure mass re- 
sponse in laboratory colonies of Solenopsis saevissima (cf. Wilson, 
i960) was applied to a small (300-worker) colony of Tapinoma 
sessile. The tested compound was allowed to evaporate from a 15 
mm 2 film into a nest inlet into which air was being gently drawn. 
The maximum increase, if any, of workers outside the nest was then 
measured during the ensuing five minutes. Each substance was tested 
three times. Tests were spaced at least two hours apart to reduce 
habituation, and substances were presented in irregular order to pre- 
vent special sequential effects. Formic acid, a common defensive 
abdominal secretion of formicine ants, caused either no visible re- 
sponse or very mild excitement in the nest, and increased the outside 
worker force by only 3-8. Concentrated Dufour’s gland secretion 
of Solenopsis saevissima , which induced a mass exodus of saevissima 
