WikiJournal Preprints/Moisture Content as a Proximate Factor in Nest Site Attractiveness for Temnothorax rugatulus

Introduction
Species of the genus Temnothorax are valuable subjects for colony emigration research. Temnothorax rugatulus colonies typically nests in small natural crevices such as hollowed seeds and branches or rock cracks, often comprising less than several hundred individuals. When nests are disturbed, colonies are adept at emigrating to new nests. Disturbances arise from a number of proximate factors, many of which likely play a critical role in collective nest site decision making. These factors may include threats such as predation; elemental risks such as wind, water, light, or heat; and conspecific or heterospecific competition. These threats and risks correspond directly to preferences of small nest entrances, small and thin nest cavities, and low light conditions. Other proximate factors may include proximity to food or water sources as a means for survival, or moisture content of a nest site to reduce desiccation, as insects are often susceptible to this condition.

These various nest-attractiveness factors may influence a colony's decision to emigrate. It has been shown that T. curvispinosus can collectively compare qualities of nest sites, distinguishing between different proximate factors. Other studies have manipulated proximate factors differentiated by nest cavity size, cavity thickness, entrance size, and light levels in Temnothorax. It has also been shown that Temnothorax will select nests that are closer to food sources; however, little work has been done when considering water.

This paper examines moisture content as a single proximate factor in nest site attractiveness. As water plays a central role in the survival of individuals, a colony must necessitate proximity to water for two reasons: to consume and to minimize desiccation risk. Here, I investigate T. rugatulus' emigration preferences to low-quality nests with differing moisture contents.

Methods
Three experiments were conducted to test the hypothesis that T. rugatulus prefers nests with high moisture content. The experiments allowed for colonies of T. rugatulus to emigrate to alternate nests of their choice after disturbance or removal of their "home" nest. The experiments were run on different days and structured like the following:

Subjects and design
Eleven colonies of T. rugatulus were housed in an artificial nest made of 2.4 mm thick balsa wood slats. Each rectangular nest (50 x 75 mm) contained a 38 mm diameter circular cutout sandwiched between two 50 x 75 mm glass microscope slides. The entrance of the nests were small, 2 mm cut notches. Before and between experiments, the colony’s housings were placed within their "home base" (11 x 11 cm plastic arenas coated in ethylene tetrafluoroethylene to prevent escape). When located in their "home base", they were provided water and an agar-based protein-rich diet based on Bhatkar & Whitcomb (1970).

The experiments were conducted in 25 x 25 cm plastic arenas, the edges coated with ethylene tetrafluoroethylene to prevent escape. Each arena contained two identical empty nests placed on the left and right sides. These nests consisted of two 2.4 mm think balsa wood slats cut into 50 x 13 mm rectangles that acted as legs to hold a 50 x 75 mm glass microscope slide. Underneath these nests was placed a thin rectangular slice of hydrophilic foam (roughly 7 x 24 mm, with a thickness of 2.4 mm or less).

After the emigration timeframe was completed, photographs were taken off each nest and the number of ants at each nest were counted and recorded (the "home" nest, the wet hydrophilic foam nest, and the dry hydrophilic foam nest). The presence of brood and queens were noted as well.

Experiment A: Nest disturbance
Two trials of n=10 colonies each were conducted over a 24 hour period to allow for emigration. Ten colonies were placed into the experimental arenas and the top slides of the "home" nests removed. Under each alternate nest choice, one side had wet hydrophilic foam and the other dry. Each arena alternated the position of the wet foam from A (left) to B (right). The lids were placed on top of the arenas and stacked. The stack was then surrounded by a cover with the top and bottom open. This was done to reduce the amount of light that could potentially interfere with nest choice decision, but was not eliminated, as T. rugatulus relies heavily on visual sense for navigation. The colonies experienced standard laboratory florescent lighting along with minimal ambient sunlight. The alternating wet foam location was designed to account for potential light gradients, as T. rugatulus will invariably choose nest sites with lower light conditions. For trial one, after 24 hours, the number of ants could be counted at each nest ("home", wet, and dry) for all ten colonies. For trial two, after 24 hours, the number of ants could be counted at each nest for all ten colonies. Experiencing a disturbed nest, under an equal likelihood probability, T. rugatulus will choose to emigrate to the nest with the highest moisture content, even if it is of low-quality.

Experiment B: Nest removal
Under the same conditions, n=9 colonies were forced to choose a nest. This was done by collecting the entire colony and brood with a pooter. The "home" nest was not placed in the arena. The collected colony was extracted from the pooter collector and placed into the arena in the space that would normally house the "home" nest. When forcing a nesting decision (by complete nest removal) to emigrate to either option of two equally low-quality nests, T. rugatulus will choose a nest with high moisture content.

A control to determine statistical significance in nest preference was conducted by replicating the previous experiment except with n=10 colonies and both nest options containing wet foam. The expected outcome here was that T. rugatulus will choose nests A and B with an equal distribution.

Statistical Analysis
The preference criteria to determine colony nest choice consisted of calculating the percentage of ants located at the wet nest out of the total number of ants present at each nest. For experiment B, the "home" nest was not present, so preference was determined by calculating the percentage of ants present at the wet nest over the total number of ants in the arena. Noting the presence of brood and queens was done to make judgments in cases of splits (where the colony chooses both nests, or even all three nests). A judgment was made when the number of ants at the wet nest over the total was >50%.

Preference criteria were set to determine judgment for purposes of creating a binary “yes or no” wet decision. These preferences were input in a chi square goodness-of-fit contingency table with equal likelihood distributions. To find a statistical preference for experiment A, the chi square was calculated in conjunction with a paired, two-tailed $$t$$-test comparing the number of ants at the wet nest versus the dry nest. For experiment B, and the control, paired $$t$$-tests were used. An additional paired $$t$$-test was calculated for experiment A comparing the number of ants at the wet nest versus the number at the "home" nest. This was done to determine if there was preference for alternative nests entirely, regardless of the presence of moisture.

Results
A chi square goodness-of-fit test was calculated using the binary preference judgments in experiment A. Ten out of twenty colonies chose the wet nests and zero chose the dry nests. The expected values were ten and ten for the wet and dry respectively. This resulted in $χ^{2}$=10; df=1; $$p$$=0.002. A statistically significant result, in such that, more ants chose the nests with wet foam than with dry foam. Note that ten chose wet and zero chose dry, equalling less than the sample size of twenty. This resulted due to ten colonies not choosing either nest, but preferring to not emigrate at all. This unexpected result ($$t$$=0.74962; df=19; $$p$$=0.4581) will be discussed below. A paired $$t$$-test comparing the number of ants that preferred the wet nest versus the dry nest was found to be highly significant ($$t$$=4.27649; df=19; $$p$$=0.000123). Experiment B also resulted in a statically significant difference in the number of ants that preferred the wet nest versus the dry nest ($$t$$=4.43932; df=8; $$p$$=0.000412). Nine out of nine that chose to emigrate, did so in favor of the wet nest, with one colony not emigrating to any nest at all. The control for experiment B was analyzed with a paired $$t$$-test, finding no significance in preference for nest A or nest B ($$t$$=-1.23442; df=9; $$p$$=0.23292).

Discussion
Experiment A lends support to the prediction, in such that, all of the colonies that chose to emigrate, did so in favor of the wet nest (ten out of ten, with one favoring the dry nests). However, upon completion of both trials, an unexpected result arose. Many of the ants did not chose either the low-quality wet nests or dry nests, but instead remained in their "home" colony despite missing a “roof”. The colonies that remained, were exposed and lacked the typical crevice-like nest they typically prefer. Upon counting the individuals that remained, a $$t$$-test was performed comparing the number of ants present at the wet nests versus the "home" nest ($$t$$=0.74962; df=19; $$p$$=0.4581). There was no statistical difference in nest preference for their "home" nest or the wet nest. As discussed previously, the total that chose nests equaled less than the sample size of 20 colonies (in binary, ten chose the wet nest, zero chose the dry nest, and ten chose their original "home" nest). This result suggests that the nest choices may have been of low enough quality to compel T. rugatulus to not emigrate. For this reason, experiment B (and its control) was devised in conjunction, simply to see if, when forced, T. rugatulus would move at all. This contributed to the confidence that the conditions of the arenas, lighting, or other unknown factors were not inhibiting colony emmigration.

Furthermore, this result prompted the scoring of the presence of queens. When identified, all colonies (out of the 20) that did not emigrate to a suitable nest, contained at least one queen. This may support the conclusions of Doering and Pratt (2016), that split colonies of T. rugatulus will often reunite at a nest that contains the queen. In this study, continual observation or recording was not done, so it is possible that in a few cases, during the 24 hour period allowed for emigration, some of the colony emigrated to one of the nests and subsequently emigrated back to the "home" nest due to the presence of the queen.

Experiment B supports the prediction that T. rugatulus, when faced with a forced decision between two low-quality nests, will choose the nest with the highest moisture content. This was supported by both the statistically significant $$p$$-value when comparing the number of ants at each nest and by the fact that 100 percent of colonies that chose to emigrate (eight out of a sample size of nine), emigrated to the nest housing the wet foam. The one colony that did not emigrate, despite the absence of a "home" nest, was a colony of less than 25 individuals, and as such, are granted greater flexibility in nest choice as their nest site can be extremely small. In this particular case, 18 individuals emigrated to the corner of the arena. Furthermore, the control for experiment B supported the prediction of an insignificant $$p$$-value, in such that, colonies did not show a specific preference for either nest A or B—establishing at A, B, both locations (colony split), or no location at all. This result was expected from the control, as both nests were of equal low-quality and both contained wet foams.

These results contrast with Karlik et al. (2016) who showed that the related species T. curvispinosus showed no preference for moisture in tests of colony organization and position within a nest. This also contrasts to the idea by Hood and Tschinkel (1990) that above ground ants are more tolerant to drier conditions than those of soil-living ants (in Karlik et al. 2016). T. rugatulus appears to favor high-moisture nests despite living outside of soil. Furthermore, this suggests that T. rugatulus may have a greater sensitivity to desiccation than previously thought. This may also correspond to the established preferences (in T. albipennis) for highly compacted spaces and small nest entrances, as these factors have been suggested to increase defense abilities, decrease predation, and allow for space to tend to their brood. In T. rugatulus, this same nest site characterization may act to confer a fitness advantage, as a compact space will naturally increase the humidity within, thereby decreasing the risk of desiccation.