1969] 
Hangartner — Solenopsis 
63 
This result supports the C 0 2 -hypothesis, but is not absolutely con- 
clusive, because Sodium hydroxide could absorb other volatiles be- 
side C 0 2 . In an attempt to get a more direct proof, I performed the 
following experiment: 5 bottles were prepared in the way shown 
in Figure 1 except that they did not contain any ants. They were 
then connected with U-shaped pieces of glass tubing (i.d. 1.5 mm) 
as demonstrated on top of Figure 3. After the whole system has 
been placed on a Wilson nest, a slow C 0 2 -stream (3.5 cm 3 /min) 
was pressed into one end of the tubing. Due to loss of C 0 2 through 
the cardboard and perhaps tiny leaks in the stoppers, only traces of 
this gas left the opening at the other end. I, thereby, got a more or 
less continuous gradient from a relatively high C 0 2 -concentration 
(vial 1) to a relatively low concentration (vial 5). Simultaneously 
I employed a second system of the same design, but compressed air 
was used instead of C 0 2 (control). In 20 hour-intervals, the card- 
board on each stopper was replaced by a new one. The results of 
five repetitions can be taken from Figure 3 (bottom). 
It is evident that a relatively low concentration of C 0 2 is able 
to release the same behavior pattern as do trapped workers. How- 
ever, the efficiency of this releaser decreases with increasing concen- 
tration. It might be of interest to notice that ants which punctured 
the cardboard of vial 3 died or were at least anesthetized after pene- 
tration into the interior of the stopper, whereas this effect could not 
Figure 3. Arrangement (top) and results (bottom) of the experiment 
to test the effectiveness of C0 2 as a releaser for digging behavior. The 
density of the dashes represents the relative concentration of carbon dioxide 
inside the vials,, and the arrows show the direction of the gas flow. The 
radius of the stippled circle areas represents the average digging response 
and the dashed circles indicate the range of five experiments. 
