1962] 
Ghent and Gary — -Alarm Rcleaser 
3 
of a fresh sting on an introduced object increases the probability of its 
being stung. Captured workers were first induced to sting a disk of 
suede leather 1.5 cm. in diameter. This was lowered on a wire into 
Table 1 . Response to suede leather disk bearing three fresh stings compared 
to response to an untreated control disk. 
(a) Disk Bearing Stings (b) Control Disk 
No. of bees No. of addi- No. of bees No. of addi- 
clustered on disk tional stings clustered on disk tional stings 
Test #1 
50 
4 
0 
0 
#2 
12 
3 
0 
0 
#3 
10 
0 
1 
0 
#4 
5 
1 
0 
0 
#5 
100 
3 
0 
0 
#6 
30 
1 
0 
0 
Test of significance for difference in number of additional stings: 
x 2 — 12 with 1 degree of freedom 
P < .001 
an observation hive and held about 5 cm. above the comb for one 
minute, then carefully removed. Allowing 30 seconds between intro- 
ductions, this disk and other control disks bearing no stings were 
alternately introduced in the same manner. The number of bees 
which congregated on the disks during each test was estimated and 
the newly deposited stings were counted after each removal. The data 
in table ( 1 ) clearly exhibit greater attractiveness of the leather disk 
bearing stings, as well as a significantly (P < .001) greater number 
of stings retained in it. 
These results seemed to indicate that the presence of a fresh sting 
alone is sufficient to provoke stinging attacks. In subsequent observa- 
tions, however, it was demonstrated that one or more secondary stimuli 
are usually necessary to elicit stinging. Although stings invariably 
attract bees, we observed very few cases of stinging unless the test 
object was moved or jarred. Indeed, bees that congregated about an 
introduced sting tended to fly at any moving object in the vicinity. 
Since the loss of the sting is often fatal to the worker, there is an 
obvious selective pressure against the stinging of immobile objects, 
even though coated with an alarm releaser. 
In subsequent experiments, it was found that extracts of stings 
attracted worker bees, and induced a pattern of behavior which was 
apparently identical to that of workers exposed to fresh stings. 
The alarm reaction was further characterized in an experiment 
relating the quantity of extract to the number of bees attracted to it. 
Th ree hundred whole uncrushed stings, pulled from freshly-frozen 
workers, were extracted in 1 ml. of methylene chloride. Measured 
