Ant Trail System
The trail system of E. procerawas investigated during March and
April 2008 at the same field site as described above. We deployed
mushroom baits (Pleurotis sp.) every 10 m on the forest floor along
existing hiking trails to attract E. procera workers. Foraging trails
were back-tracked to the nest site wherever possible. Nests were
then marked for further observations. Since foraging of E. procera
was found to be strictly nocturnal (Witte & Maschwitz, 2008), observations
took place between 2000 and 0400 hours. Occasional
activity measurements were also taken during daytime (1000e
1700 hours) to verify the nocturnal lifestyle.
To analyse the foraging pattern of E. procera in detail, we
recorded entire trail systems of five randomly chosen colonies over
a period of 32 nights (for details on how long each colony was
observed see Fig. 2 and Supplementary Fig. S3). On 10 nights no
data could be collected because of heavy rain. On the first few
nights of observation, we placed small flags (wooden sticks with
tape) into the ground every 50 cm along all active trails. Activity at
trails was subsequently surveyed every night between 2000 and
0400 hours. To measure activity, we counted ants passing the first
flag outbound from the nest. If a minimum of five ants per min
passed the first flag, the route was considered as being active, and
deviations from all previously installed flags (first day distances,
FDDs) or from the trail of the previous day (previous day distances,
PDDs) were measured on the respective trail (Fig. 3). Deviations on
an active trail were measured only once per night. While FDDs
provide information on the consistency of trails over a long period
of time, PDDs tell us whether trails were accurately followed from
one day to the next. Distances were measured in centimetres. From
10 cm on, distances were measured in 5 cm increments. New flags
were installed when ants established a yet unrecorded trail from
the nest site, or new side branches from or extensions of former
trails. A trail was considered new if three sequential flags showed
deviations of more than 5 cm (for explanation of trail system
recording see Fig. 3). Furthermore, the route system of one colony
was mapped at real scale using a grid of strings (16 m 16 m) with
a mesh width of 1 m (Fig. 2), including trees, shrubby plants and
dead branches lying on the ground. In addition, the route system of
four colonies was drawn on a spatial sketch of the environment at
the end of the observation period. The growth locations of mushroom
patches were mapped as well, although not as systematically
as in the separate study on mushroom characteristics.
General ant foraging activity was measured by carefully
inspecting the nest entrance surroundings of 14 colonies, including
the five used for studying the route system. First, we noted whether
ants were present or not. If ants were present, traffic flow on trails
was measured by counting both inbound and outbound ants
passing an imaginary line perpendicular to the trail about 50 cm
from the nest.