1984] 
Holldobler — New exocrine gland 
233 
Fig. 6. SEM-photograph of the pygidium of a worker belonging to the Formica 
sanguinea group. Note: no cup structure on the anterior margin of the 7th abdominal 
tergite. 
around the leader ant, often moving briefly ahead, but always 
returning to the leader. Myrmecologists have long noted this pecu- 
liar looping behavior at the front of the raiding columns of Poly- 
ergus (see for example Schmitz 1906), but only the recent studies by 
Topoff have indicated that this appears to be due to the orgainizing 
role and excitement emanating from the leader ant. These observa- 
tions suggest that the leader discharges not only a longer lasting trail 
pheromone but also an arousal signal, possibly from the pygidial 
gland. 
Similar recruitment patterns have been observed in Camponotus 
socius and Myrmecocystus mimicus, where secretions from the poi- 
son gland seem to function as arousal signals (Holldobler 1971, 
1981), as well as in Camponotus ephippium, where the arousal sig- 
nal appears to originate from the cloacal gland (Holldobler 1982). 
However, this hypothesis cannot be applied to the well developed 
pygidial gland in Polyergus queens. Wasmann (1915) first proposed 
that a young Polyergus queen, in order to found a new colony, has 
to intrude a colony of a slave ant species, where she kills the resident 
