Introduction

Biogenic amines are evolutionarily conserved neurochemicals that function in invertebrates and vertebrates as neurotransmitters, neurohormones and neuromodulators (Libersat and Pflueger 2004; Huber 2005; Farooqui 2007; Neckameyer and Leal 2009). In insects, monoamines modulate the neural circuitry of a broad spectrum of behaviors including intraspecific aggression (Stevenson et al. 2005; Dierick and Greenspan 2007; Rillich et al. 2011), olfactory sensitivity and odor discrimination (Dacks et al. 2006; Galizia and Rössler 2010), circadian rhythms (Yuan et al. 2005), memory formation (Schwaerzel et al. 2003), and behavioral state transitions in species exhibiting life-history polyphenisms (Anstey et al. 2009). In social insects, worker monoamine amine levels can vary contextually and are associated with behavioral development and division of labor. Octopamine (OA), for example, has robust effects on foraging activity in the honey bee Apis mellifera: OA levels increase in the brains of honey bee workers as they age and gain experience outside the nest (Schulz and Robinson 1999; Wagener-Hulme et al. 1999; Schulz et al. 2003). Experimental elevation of OA causes young bees to depart the nest earlier to initiate foraging (Schulz et al. 2002a, b), alters resource selection by foragers (Giray et al. 2007), and affects how foraging-site profitability is represented by workers through dance language (Barron et al. 2007). OA also has neuromodulatory effects on within-nest honey bee behavior, including hygienic activities (Spivak et al. 2003). In addition to OA, brain serotonin (5HT) and dopamine (DA) levels in honey bees vary with development and differ among workers serving distinct behavioral roles (Taylor et al. 1992; Bozic and Woodring 1998; Kirchhof et al. 1999; Schulz and Robinson 1999), suggesting the aminergic modulation of honey bee temporal polyethism is not restricted to OA.

Ants are eusocial hymenopterans convergent with honey bees in important aspects of colony organization. OA, DA, and 5HT have been implicated in ant behaviors such as nestmate recognition (OA: Vander Meer et al. 2008), social dominance (OA and DA: Cuvillier-Hot and Lenoir 2006), trophallaxis and social interaction (OA and DA: Boulay et al. 2000; Wada-Katsumata et al. 2011), aggression (5HT: Kostowski et al. 1975; 5HT and DA: Hoyer et al. 2005), and polyethism (OA: Wnuk et al. 2011; 5HT: Seid et al. 2008; 5HT and DA: Seid and Traniello 2005), suggesting these neurochemicals modulate circuitry underscoring worker task performance in ants as well as bees. Seid and Traniello (2005) examined the relationship between task performance and amines in Pheidole dentata by measuring 5HT, DA and OA titers in individual brains in minor workers in four distinct age cohorts. Levels of DA and 5HT increased significantly from eclosion to ca. 16–20 days, coincident with the temporal structure of repertoire expansion and attainment of task efficiency (Seid and Traniello 2006; Muscedere et al. 2009), as well as the maturation of the serotoninergic system (Seid et al. 2008). The pattern of 5HT increase was striking: across the four age cohorts, brain 5HT levels increased 2- to 3-fold, with a marked increase in 5HT between age classes 3 and 4, the developmental time at which minor workers add extranidal activities such as foraging to their existing repertoire of within-nest behaviors. Seid and Traniello (2005) and Seid et al. (2008) thus hypothesized that increasing 5HT titer modulates perceptual capabilities associated with foraging activity in P. dentata minor workers. Because foraging involves chemically mediated recruitment and orientation, we predicted that 5HT would affect worker responsiveness to olfactory social signals that provide orientation guides to food sources.

To test the hypothesis that 5HT plays a neuromodulatory role in responsiveness to chemical stimuli controlling foraging, we experimentally altered brain 5HT levels in mature P. dentata minor workers and determined the ability of 5HT-depleted individuals to respond to and accurately follow pheromone trails, which are social signals that provide primary olfactory information necessary for the organization of cooperative foraging and forager orientation. Because young minor workers, which are characterized by relatively low brain titers of 5HT levels, rarely forage in nature or in laboratory colonies, we hypothesized that mature minors with experimentally lowered 5HT would show decreased olfactory responsiveness to chemical trails relative to that of nestmates with unmanipulated 5HT levels. We tested this hypothesis by pharmacologically altering brain 5HT levels in mature minor workers and recording their response to artificial pheromone trails.

Methods

Colony maintenance, experimental subcolony establishment, and 5HT depletion

Queenright colonies of P. dentata were collected in Florida and reared in the laboratory as described in Muscedere et al. (2009). To standardize the social environment of drug-treated and control workers, we constructed replicate subcolonies composed of 45 mature, fully pigmented minor workers [age class (AC) 4, Seid and Traniello 2006; Muscedere et al. 2009], 5 major workers, 5 late-instar larvae, and 5 young pupae. Each subcolony nested in a foil-covered test tube plugged with moistened cotton. Control subcolonies were provisioned with 100 μl of 2 M sucrose solution; experimental subcolonies received 100 μl of 2 M sucrose with 4 mg/ml of α-methyltryptophan (AMTP), which selectively inhibits 5HT synthesis in insects (Sloley and Orikasa 1988; Stevenson et al. 2005). Subcolonies received no other food for the duration of the experiment. Subcolonies were fed on three schedules: on days 0, 2, 4, and 6 (assayed on day 7); on days 0, 2, 4, and 7 (assayed on day 8); and on days 0, 2, 4, 6, and 9 (assayed on day 10). No behavioral differences were observed among workers experiencing different feeding regimes and there was no detectable significant effect of feeding regime on any trail-following metric recorded and analyzed. Data were therefore pooled for analysis. Five queenright laboratory stock colonies were used as sources of workers for all control and experimental subcolonies. Seven to 12 minor workers were assayed from each of the five stock colonies (total N = 40, 20 minors per treatment).

Quantifying olfactory responsiveness to trail pheromone

Minor workers from control and AMTP-treated subcolonies were exposed to artificial pheromone trails of standardized concentration to assess trail-following response. We focused on minor workers because they are numerically dominant in colony populations and perform the majority of labor, including foraging. “Worker” hereafter thus refers exclusively to minor workers. To extract trail pheromone, poison glands [the source of trail substance in Pheidole (Hölldobler and Wilson 1990)] were dissected from mature minor workers collected from colonies not used as sources of experimental individuals. Glands were homogenized in 100% ethanol (1 gland/5 μl ethanol). We then evenly applied 10 μl of trail extract with a microsyringe (Hamilton Co. #701) along a circular trail 16 cm in circumference lined in pencil on filter paper. We used concentrated trails (2 glands per 16 cm trail) to ensure workers were exposed to pheromone concentrations exceeding their threshold of sensory detection, which we estimate at 0.02–0.2 glands per 16 cm trail (unpublished data). Focal workers to be assayed were removed singly from subcolonies, briefly (ca. 20–30 s) anesthetized on ice, then gently moved to the center of a circular artificial trail in an open petri dish arena with walls coated with Fluon® to prevent escape. After ants recovered from anesthesia (~30 s after placement in the arena), individuals were digitally videotaped with a Sony DSC-H5 camera for 10 min as they freely explored the arena and interacted with the artificial trail. Each individual was assayed on a freshly deposited trail on a fresh piece of filter paper to ensure workers were tested with trails of comparable age and pheromone concentration. Each videotaped trial was then analyzed using JWatcher Video v1.0 (http://www.jwatcher.ucla.edu) to record the number of encounters with a trail and the number, length and duration of each trail-following event. We defined a trail encounter as movement within one minor worker body length (approximately 5 mm) of an artificial trail, and the end of an encounter as movement more than one body length away from the trail. For each trail encounter, ants responded positively or negatively. We defined a positive response as orientation along the trail axis and tropotactical movement for a distance of at least 5 mm along the trail axis. We defined a negative response as worker movement across the trail without changing direction or actively moving away from the trail without initiating following. For each positive trail-following response, we recorded the length (estimated to the nearest 5 mm) and duration of orientation along the trail. We then calculated summary statistics for each worker assayed, including average distance traveled per encounter (sum of all distances followed divided by total number of trail encounters), average distance traveled per positive response (sum of all distances followed divided by total number of positive trail-following responses), and average velocity during trail following in cm/s (sum of all distances followed divided by total time following trails). We also noted the single longest positive following response per trial and the proportion of trail encounters that were negative responses for each worker. We did not include data from assays in which a focal worker encountered the trail less than 5 times. One AMTP-treated worker encountered the artificial trail more than 5 times but never followed it; by definition, this trial was included, accounting for sample size differences (N 1 = 19 or 20) in several statistical tests. For example, this worker was assigned a value of 1.0 for proportion of encounters eliciting a negative response, but no value for average distance traveled per positive response.

Measurement of brain 5HT and DA content

Immediately after each trail-following assay, minor worker brains were prepared for analysis of 5HT content using isocratic, reversed-phase high-performance liquid chromatography with electrochemical detection (HPLC-ED). The HPLC-ED system used included the following components, all manufactured by ESA, Inc (Chelmsford, MA): a model 584 pump, MD-150 (3 × 150 mm) reversed-phase analytical column, a 5011A dual-channel coulometric analytical cell, and a Coulochem III electrochemical detector. We measured 5HT and DA levels in all the AMTP-treated and control mature workers assayed for trail-following ability (N = 40), as well as an additional sample of unmanipulated, newly eclosed workers [N = 11; age class (AC) 1, Seid and Traniello 2006; Muscedere et al. 2009] taken directly from stock colonies, to compare naturally occurring differences in 5HT and DA among worker age cohorts. All brains were quickly dissected in cold Ringer’s solution, individually homogenized in 55 μl mobile phase, centrifuged for 10 min at 15,000 rpm at 0°C, and kept on ice in the dark to prevent amine degradation if not analyzed immediately. Mobile phase chemistry (50 mM citrate/acetate buffer, 1.5 mM sodium dodecyl sulfonate, 0.01% triethylamine, and 24% acetonitrile in MilliQ water) was modified from Hardie and Hirsh (2006). All samples were analyzed on the day of dissection. Electrode potentials were set to −125 and 225 mV for the first and second channels, respectively. Twenty micro liter of supernatant from each sample was injected onto the HPLC column and levels of amines were detected on channel two. External standards [serial dilutions of 5HT and DA (Sigma-Aldrich) in mobile phase] were run each day, to quantify amine levels.

Although 5HT, DA and OA have been detected in P. dentata minor worker brains (Seid and Traniello 2005), and OA and DA have been implicated as modulators of ant behaviors (see references cited in “Introduction”), we focused on the association between 5HT levels and trail-following behavior due to the striking increase in total 5HT content (Seid and Traniello 2005) and number of 5HT-immunoreactive neurons (Seid et al. 2008) in the maturing P. dentata minor worker brain. Because simultaneous, highly sensitive HPLC-ED quantification of 5HT, DA, and OA in minute Pheidole brains is difficult due to the high detector potentials needed to oxidize OA, we optimized our HPLC-ED method for the selective and accurate quantification of 5HT. DA, but not OA, can be detected incidentally using the same method. We therefore comparatively report DA levels in our experimental animals to illustrate that AMTP treatment selectively lowers 5HT but not DA levels (see “Results”).

Statistical analyses

Differences among worker groups in brain 5HT and DA content were assessed with ANOVA. When analyzing DA levels, we excluded two strongly positive outlying DA samples from the control mature workers group because we were confident they were contaminated with minute fragments of cuticle, which contains substantial quantities of DA but no 5HT. Exclusion of these samples is justified because both had extraordinarily elevated DA (>3 standard deviations from mean DA content) but normal 5HT levels. Nonetheless, inclusion of these two outliers would not have altered the significance of the reported tests. The distributions of trail-following metrics tended to be non-normal; differences among treatments on these measures were therefore assessed with Mann–Whitney U tests, and quantitative relationships between variables were assessed using Spearman’s rank correlations.

We performed sensitivity analyses to determine if experimental or methodological biases in our behavioral observations could have accounted for differences observed between AMTP-treated and control workers. We performed three sets of tests in addition to the main analyses described above. First, we tested whether using a 5 mm trail-following event as our criterion for a positive response could have biased our results due to the error inherent in differentiating between 5 mm-following events and negative responses. To do so, we recoded the original data set used to conduct the initial analyses by changing all recorded 5 mm-following responses to negative responses. By doing so, the criterion for a positive response effectively became 1 cm, which is unambiguously a positive response. Second, we tested the subjectivity of our criteria by having an entirely new data set generated by a naïve, second observer uninvolved in the original experiments and blind to the experimental rearing treatment of each worker. This observer independently scored behavior from all videotaped trials, using a criterion of 1 cm to discriminate a positive response. Third, we averaged the recoded values of the first observer (using a 1 cm criterion for positive response) and the second observer for each experimental worker. For all three of these new data sets, we repeated statistical tests comparing AMTP-treated and control workers using the three metrics that significantly differed in our original analyses [average distance travelled per encounter, average distance traveled per positive response, and proportion of negative responses (see “Results”)].

Results

5HT brain titers in drug-treated workers

5HT titer (ANOVA: F 2,48 = 60.2, p < 0.0001) and DA titer (ANOVA: F 2,46 = 60.2, p < 0.0001) significantly differed among unmanipulated young workers, mature control workers, and mature workers treated orally with AMTP (Fig. 1). Among mature subcolony workers assessed for trail-following ability, AMTP-treated workers had significantly lower 5HT titers, but did not have significantly altered DA titers (Tukey–Kramer post hoc comparisons, Fig. 1), indicating AMTP treatment selectively depleted brain 5HT content. AMTP treatment reduced 5HT titer approximately 43% (mean ± SEM: control, 57 ± 1 pg/brain; AMTP-treated, 33 ± 3 pg/brain). While significantly reduced, the 5HT content of AMTP-treated worker brains was on average similar to that of newly eclosed young minor workers (Fig. 1) and thus was in the range of 5HT levels typical of normally developing individuals.

Fig. 1
figure 1

Oral administration of AMTP selectively lowers brain 5HT levels in P. dentata minor workers. Mean ± SEM brain 5HT (gray bars) and DA (white bars) levels are shown for mature workers used in trail-following assays (treated with AMTP or control solutions, as described in the text), and unmanipulated young workers removed directly from stock colonies, shown for comparison. Levels that do not share letters are significantly different (Tukey–Kramer post hoc comparisons)

Full size image

Behavioral responsiveness to trail pheromone

AMTP-treated mature workers showed reduced trail-following behavior in comparison to control mature workers (Fig. 2). The average distance traveled per trail encounter was significantly lower for AMTP-treated workers (Fig. 2a; Mann–Whitney U test: N 1 = N 2 = 20, U = 102.5, p = 0.007). The significant reduction in average distance traveled per trail encounter for AMTP-treated workers is due to the fact that when AMTP-treated workers responded positively to trails, their average distance traveled was significantly shorter than that of control workers (Fig. 2b; N 1 = 19, N 2 = 20, U = 117, p = 0.041), and their proportion of negative responses per trial was significantly higher (Fig. 2c; N 1 = N 2 = 20, U = 113, p = 0.018). Thus, AMPT-treated workers were both less likely to initiate trail following and, when they did follow trails, they followed trails for shorter average lengths. We reasoned that if the effect of AMTP-treatment on trail-following behavior was in fact mediated by 5HT-depletion, then there should be a direct relationship between worker 5HT titer and trail-following performance independent of worker experimental treatment. Across the entire dataset, brain 5HT titer and average distance travelled per trail encounter were significantly correlated (Fig. 3; Spearman’s correlation: ρ 38 = 0.35, p = 0.028).

Fig. 2
figure 2

Trail following is impaired in mature P. dentata workers pharmacologically depleted of brain 5HT. Box plots show median values (horizontal lines), interquartile ranges (boxes), and 1.5 × interquartile ranges (whiskers). AMTP-treated workers follow trails significantly shorter distances than control workers a per encounter, which is a result of b following less per positive response, and c responding less frequently to trails when encountered

Full size image
Fig. 3
figure 3

Worker trail following performance is significantly correlated with brain 5HT content. Open symbols represent AMTP-treated workers, closed symbols represent control workers

Full size image

Values of the single greatest trail length followed per trial did not significantly differ between AMTP-treated and control workers (N 1 = 20, N 2 = 20, U = 138.5, P = 0.096). Average distances followed per encounter were relatively low compared to the distance over which workers might be expected to forage in nature [medians: 17.0 mm (control), 14.6 mm (AMTP-treated)], but workers occasionally oriented along trails for long distances. The greatest trail length followed per trial thus tended to be considerably longer than the average distance traveled per encounter [medians: 75.0 mm (control), 60.0 mm (AMTP-treated); longest single length followed by any worker in each treatment: 515 mm (control), 300 mm (AMTP-treated)].

Worker velocity during trail following was positively correlated with average distance travelled per encounter (ρ 37 = 0.52, p = 0.0006), average distance travelled per positive response (ρ 37 = 0.58, p < 0.0001), and greatest trail length followed per trial (ρ 37 = 0.57, p < 0.0001), but was not correlated with the proportion of negative responses per trial (ρ 37 = −0.16, p = 0.32). However, running speed during trail following did not significantly differ between control and AMTP-treated workers (N 1 = 19, N 2 = 20, U = 144, p = 0.2; medians: control, 1.1 cm/s; AMTP-treated, 0.89 cm/s), nor was running speed significantly correlated with brain 5HT content (ρ 37 = 0.31, p = 0.06).

Statistical sensitivity analysis

Our sensitivity analyses confirmed that, regardless of the experimenter viewing videotaped trials or the criterion used to discriminate positive from negative responses (5 mm or 1 cm), AMTP-treated workers traveled significantly shorter distances per trail encounter than control workers (Table 1). In our reanalysis of the original dataset using a criterion of 1 cm for a positive response (5 mm-positive responses recoded to negative responses), average distance traveled per trail encounter was significantly lower for AMTP-treated workers (N 1 = N 2 = 20, U = 105, p = 0.009). In this analysis there was no significant difference between AMTP-treated and control workers in average distance traveled per positive response (N 1 = 19, N 2 = 20, U = 137, p = 0.14), but AMTP-treated workers still had significantly higher proportions of negative responses per trial than control workers (N 1 = N 2 = 20, U = 97.5, p = 0.005). Analysis of the blind observer’s dataset confirmed that average distance travelled per trail encounter was significantly shorter for AMTP-treated workers (N 1 = N 2 = 20, U = 127, p = 0.049). However, comparisons of the other metrics did not meet the threshold of statistical significance when using the second observer’s measurements (average distance traveled per positive response: N 1 = 19, N 2 = 20, U = 143.5, p = 0.29; proportion of negative responses, N 1 = N 2 = 20, U = 144.5, p = 0.13). Finally, analysis of the averaged values recorded by the first and second observers again indicated that AMTP-treated workers travelled significantly shorter distances per trail encounter than control workers (N 1 = N 2 = 20, U = 107, p = 0.011). In this analysis, average distance travelled per positive response did not significantly differ between treatments (N 1 = 18, N 2 = 20, U = 142, p = 0.28), but the proportion of negative responses per trial was significantly higher for AMTP-treated workers (N 1 = N 2 = 20, U = 113, p = 0.018).

Table 1 Summary of statistical comparisons of trail-following metrics
Full size table

Discussion

Behaviorally mature P. dentata minor workers attend to foraging and defense through pheromonally organized systems of recruitment communication and trail orientation. Pheromone trail constituents induce recruitment behavior and provide orientation cues for workers engaged in food retrieval and defense, maintenance of territory, and nest security. Previous research demonstrating that only fully pigmented, developmentally mature minor workers are active in performing tasks outside the nest indicates worker responsiveness to social signals of foraging and defense is age-related (Wilson 1976; Seid and Traniello 2006). Although mature minor workers attend to foraging and defense as well as within-nest tasks such as nursing, relatively young workers lack this behavioral totipotency, apparently due to physiological and anatomical immaturity (Seid and Traniello 2006; Muscedere et al. 2009, 2011). Our present study contributes to this work in two ways. First, we confirmed the analysis of DA and 5HT levels presented by Seid and Traniello (2005): among workers not treated with drugs, DA levels were higher than 5HT levels, and differences in 5HT and DA titer between young and mature workers (AC1 and AC4 workers, respectively) were pronounced (ca. 2- to 3-fold). Absolute values of amine levels reported in the two studies differ, however, likely due to the use of different HPLC systems and variation in methods. Second, we experimentally demonstrated that 5HT acts as a neuromodulator of responsiveness to recruitment chemicals and trail orientation in mature minor workers, providing the first demonstration of the neuromodulation of trail communication in ants. For each encounter with an artificial trail prepared from poison gland contents, minor workers with pharmacologically lowered brain 5HT levels traveled shorter distances on average than control workers with unaltered levels of 5HT. The present study investigated only how brain 5HT levels affect worker trail-following performance. Future studies will explore how P. dentata worker behavior is affected by monoamines such as DA and OA (Libersat and Pflueger 2004).

Serotonergic neurons are present in all compartments of the ant brain (Tsuji et al. 2007), including regions of motor control (the central body), primary olfactory input (the antennal lobes) and higher-order sensory processing (the mushroom bodies). 5HT immunoreactivity is evident throughout the P. dentata minor worker brain, including in the antennal lobes, where it is characterized by fine varicosities often encircling and innervating olfactory glomeruli, and in the mushroom body calyces, which are innervated by olfactory projection neurons from the antennal lobes (Seid et al. 2008). Serotonergic somata are also prominent in the optic lobes, and significantly increase in number as workers age (Seid et al. 2008). Given the ubiquitous distribution of 5HT immunoreactivity in the P. dentata minor worker brain, the impact of inhibiting 5HT synthesis on sensory, motor and integrative brain functioning is potentially significant to task performance. 5HT depletion could cause trail-following deficits by modulating general olfactory responsiveness and/or sensitivity to specific trail pheromone components, higher-order olfactory processing and sensory integration, motivation to search for food and/or follow trails, or any combination of these effects. General inhibition of motor functions influencing motor control of locomotion and navigational abilities, including worker velocity during trail orientation, could also have contributed to the reduced trail-following performance of AMTP-treated workers. Although not significantly different, control workers tended to have higher velocity than AMTP-treated workers, and velocity was significantly correlated with average trail-following length. However, even if AMTP-treated workers had slower movement while following trails, it is unclear if this is due to a direct inhibitory effect of 5HT depletion on motor output and movement, or an indirect effect of 5HT depletion on olfactory sensitivity and/or processing, causing workers to engage in slower, more deliberate sampling of the trail active space. To test these possibilities, general movement rather than only movement rate while following trails could be examined, in addition to more detailed analysis of recruitment and orientation behavior. Nevertheless, three of our four analyses indicate a significant increase in the frequency with which 5HT-depleted workers encountered but did not follow trails, which was not correlated with worker velocity. If these results are confirmed, then we suggest the reduced trail-following responses of AMTP-treated individuals, particularly their increased likelihood of showing no response to an encounter with a trail, cannot be attributed solely to a systematic reduction in movement rate in drug-treated workers.

Despite the diversity of neuronal populations potentially affected by 5HT depletion, comparisons with other insect taxa suggest that reduced olfactory sensitivity mediated by altered functioning of serotonergic circuits in the antennal lobes could lead to the reduced trail-following behavior we observed in AMTP-treated workers. Serotonergic innervation of the antennal lobes is conserved across insect taxa (Dacks et al. 2006; Galizia and Rössler 2010), and the antennal lobe glomeruli of ants, including those of P. dentata minor workers, are innervated by serotonergic processes (Tsuji et al. 2007; Seid et al. 2008; Zube and Rössler 2008). In moths, 5HT modulates the firing rate and sensory thresholds of antennal lobe neurons in response to plant volatiles and conspecific pheromones (Gatellier et al. 2004; Dacks et al. 2008; Kloppenburg and Mercer 2008). Higher 5HT levels are associated with increased responses and lower, more sensitive thresholds. Our ongoing research is exploring whether 5HT-depleted P. dentata workers have higher olfactory thresholds for trail pheromones and other socially relevant odors (e.g., nestmate recognition pheromones, brood odors, or prey stimuli), which could indicate a similar role for 5HT in ants.

Although 1- to 2-day-old minor workers have significantly lower brain 5HT levels than mature workers (Seid and Traniello 2005), observations indicate newly eclosed minors will nevertheless follow artificial trails with high accuracy (ML Muscedere, pers. obs.). 5HT titer alone is thus inadequate to fully explain the modulation of olfactory sensitivities and motor functions related to trail-following behavior. However, young minor workers require physical prodding to induce movement and thus an encounter with an artificial trail, and their velocity during orientation is extremely slow, potentially improving their pheromone sampling (ML Muscedere, pers. obs.). These experimentally displaced young workers may be highly motivated to follow artificial trails, interpreting their chemical information as a guide to locate the nest. This behavioral flexibility could be adaptive: trail following by newly eclosed minor workers may be necessary for nest emigration after the detection of predation risk or other nest disturbances (Droual 1984). Additionally, workers of some ant species require behavioral stimulation from a successful forager to initiate trail following (Hölldobler and Wilson 1990). The circuitry underscoring trail-following behavior therefore appears to be modulated by multiple social and neurochemical signals, including 5HT. Although the modulation of trail following in ants remains to be comprehensively analyzed, our present results clearly illustrate that 5HT depletion alters the behavioral responsiveness of mature P. dentata minor workers to trail pheromone and provides experimental support for the hypothesis that task responsiveness is under aminergic regulation.