Methods
Our focal species was the thieving ant, Ectatomma ruidum . This is
one of the most common species in the Neotropics, ranging from northern
Mesoamerica down to western Amazonia, often occurring extraordinarily
high densities (Guénard & McGlynn, 2013; Santamaría, Armbrecht, &
Lachaud, 2009). This species has been characterized as thermophilic, as
it is successful in areas with ample sunlight and foraging occurs even
during the hottest times of the day (McGlynn et al., 2010). With a
generalized diet (Jandt, Hunt, & McGlynn, 2015) and moderately-sized
colonies of about 200 workers that can be collected in their entirety
(Jandt et al., 2015), it is readily maintained in captivity. Workers are
about 10 mm long, so they are large enough to individually mark,
observe, and subject to thermal assays.
Workers in E. ruidum forage outside their nests throughout a home
range that typically covers several square meters, and foragers are
typically generalists that mostly hunt for arthropods, but also collect
seeds and nectar (Jandt et al., 2015). From earlier experiments, it is
known that foragers represent a only a small fraction of the total
number of ants in the colony, and based on marking of individuals,
foragers will remain in this role for at least several weeks (Guénard &
McGlynn, 2013). While workers are observed foraging at all times of day,
it is not known if there is a division of labor between foragers that
forage diurnally and those that forage nocturnally. While foragers are
active in very hot conditions, colonies in these sunny areas will
preferentially relocate their nests underneath shade by excavating a new
nest and abandoning the old one when shade is provided (McGlynn et al.,
2010).
Work was conducted at La Selva Biological Station, located in the
Caribbean lowlands of northeastern Costa Rica (10.4306° N, 84.0070° W)
in May-June 2018. La Selva is located in a lowland tropical wet forest
that receives about four meters of rainfall annually, the bulk of which
arrives during the wet season between June and December (McDade, Bawa,
Hespenheide, & Hartshorn, 1994). We conducted work during typical
weather for the start of the wet season, with frequent but irregular
rainfall and a sharp daily increase in temperature in the afternoon.
Experiments were conducted with ant colonies located on periphery of the
laboratory clearing, located under partial canopy cover.
Our principal response variable in this study is the capability of ants
to withstand a constant challenging temperature. We measured this in a
time-to-failure assay, which we henceforth refer to as “thermal
persistence.” We chose this approach because it is most representative
of the biological phenomena we are investigating. After pilot
experiments, we chose against estimating critical thermal maxima
(CTmax) because this approach produces little
intraspecific variance in this species (Esch et al., 2017; Nelson et
al., 2018) and there is poor reproducibility (Ribeiro et al., 2012). In
the thermal persistence assay, ants were individually loaded into 2 mL
plastic microcentrifuge tubes and subjected to a constant temperate of
42ºC in a heating/cooling block (Tropicooler 260014, Boekel Scientific,
Feasterville, PA, USA). We evaluated whether the ants were capable of
maintaining a righting response upon rotation, in 1 min intervals. (The
microcentrifuge tube containing the ant was removed from the heating
block, held horizontally, and was rotated for about 3 seconds at a rate
of about 2 revolutions per second, and we assessed whether the ant
remained standing during rotation. When the ant failed to demonstrate a
righting response at the interval, the threshold time in minutes was
recorded.) Throughout this study, the persistence time before failing
the righting response ranged between 3-35 min. The same researcher (JG)
performed all assays to maintain consistency in the protocol for the
righting response assay. After trials, ants were removed from the tubes
and recovered in a lidded plastic container, and then were euthanized at
4ºC, as returning them to their nestmates would have interfered with the
validity of the experiment.
We conducted two experiments. In the first, henceforth “the marking
experiment,” we evaluated the thermal persistence of ants based on
their behavioral role in colonies, indicated by foraging activity and
position in the nest in the field. We found and marked nests of 11
colonies in the field. Over the course of a minimum of two days, we
marked foraging workers on the gaster with Testor’s Enamel (Vernon
Hills, IL), with colors to indicate to thermal maximum (early afternoon)
or thermal minimum (early in the morning). This marking technique has
been routinely used for this purpose and does not appear to interfere
with behavior (Breed, McGlynn, Stocker, & Klein, 1999; McGlynn,
Shotell, & Kelly, 2003) and our pilot trials indicated it does not
interfere with thermal persistence assays. After marking, workers were
allowed to return to their nests and we continued to mark colonies to
exhaustion (until all foraging workers were observed with a mark). We
then excavated colonies with care to collect all of the individuals and
brood in the nests. E. ruidum nests are typically composed of a
vertical series of chambers, each one separated from its neighbor by a
passageway. Colonies typically have 4-6 such chambers. We separated the
ants from the two uppermost chambers of the nest from those in the lower
chambers, to use as an additional variable of behavioral role in the
colony. Ants were housed in the laboratory in 120 mm wide cylindrical
polypropylene containers (nest boxes) with a tight-fitting polypropylene
lid in shade at ambient temperature, each colony in their own container,
provided with water and stray insects as food once in the evening. These
nest boxes were opaque and did not permit light. Within a few hours of
collection in early afternoon, we conducted the thermal persistence
assay on workers, taking care to complete trials on a single colony
within a time period of a few hours.
In the second experiment, henceforth the “circadian rhythm
experiment,” we manipulated the thermal environment of colonies to
evaluate the capacity of workers to acclimate to new thermal regimes,
and to test for the effect of a circadian rhythm. We marked 10 pairs of
colonies in the field. We then excavated one colony per pair, leaving
the other colony from each pair in the field to provide workers for
control assays run concurrently with the excavated colonies. Each
excavated colony was then divided into three equivalent colony
fragments, housing them in the same type of chambers used in the prior
experiment. For each set of colony fragments corresponding to each
colony, we randomly assigned them to one of three treatments: ambient,
ambient + 3ºC, and constant 24ºC. Over a minimum of three days, we
performed thermal persistence assays on workers in colony fragments from
all treatments at two time points in each day, at the daily thermal
minima (04:30-06:30) and daily thermal maxima (12:30-14:00). These
assays included ants from the untreated paired colony in the field as a
control for artifacts of the laboratory manipulations. Each batch of
assays included 16 ants with equal sample sizes of treatments
distributed evenly among two colonies.
All analyses were conducted in R (R Core Team, 2020), including the
“tidyverse” set of packages (Wickham et al., 2019). We used simple
linear Cox Proportional Hazard Analyses (Lin & Wei, 1989), with the
“survival” package (Therneau, 2020) to evaluate differences in
time-to-failure in the thermal persistence assays. Proportional Hazard
analyses generate a hazard ratio, in which values above 1 indicate a
lower time-to-failure relative to a reference. In this instance, low
hazard ratios indicate relatively greater capacity to persist at a
challenging temperature, and colonies high hazard indicate an earlier
failure of the thermal persistence assay. None of the samples were
censored in these analyses, as all variates represented ants that were
observed until failure to perform.