Article Comparative Quantification of Trail-Following Behavior in Pest

Ricardo J. Vázquez 1, Philip G. Koehler 2 and Roberto M. Pereira 2,* 1 Bayer Crop Science, Waterman, IL 60556, USA; [email protected] 2 Department of Entomology & Nematology, University of Florida, Gainesville, FL 32611, USA; pgk@ufl.edu * Correspondence: rpereira@ufl.edu

 Received: 14 November 2019; Accepted: 17 December 2019; Published: 19 December 2019 

Abstract: A comparison of trail-following movement parameters of six major urban pest ants, Nylanderia fulva (Forel) (Hymenoptera: Formicidae), Pheidole megacephala, Linepithema humile (Mayr), Solenopsis invicta Buren, Paratrechina longicornis (Forel), and Technomyrmex albipes (Smith) demonstrated several differences in velocity of movement, amplitude of the deviations from a marked trail, percent fidelity to the trail, number of curves per unit of trail, and trail-following accuracy. Paratrechina longicornis and N. fulva had the largest deviations from the marked trails and moved three times faster (25–30 mm/s) along the trail than the slowest , S. invicta (< 10 mm/s), with other ants following between these extremes. differences in relation to going toward or returning from food were observed in a few cases, especially with Pa. longicornis for which velocity, amplitude, and trail fidelity differed between the foraging and return trails. Quantification of ant trail-following movement parameters can be useful in understanding the mechanics of ant movement and may be important in testing specific strategies and products that disrupt trail-following behavior.

Keywords: tawny crazy ant; bigheaded ant; ; red imported fire ant; black crazy ant; white-footed ant; foraging behavior;

1. Introduction Foraging behavior has been studied in several ant species such as the Argentine ant, Linepithema humile (Mayr) [1–3], the western harvester ant, Pogonomyrmex occidentalis (Cresson) [4], and the -cutting ant octospinosus (Reich) [5] and others. When a foraging ant the nest, it follows a random or consistent path, exhibiting a higher frequency of turns further away from the nest. Once a food resource is encountered the foraging ant will communicate the quality of the food source and its location to its nestmate in an effort to recruit additional foragers [6]. Recruitment in ants includes chemical communication with other nestmates that their assistance is needed to gather resources. Ants will follow a marked trail, with movement that can be characterized by several parameters such as speed and weaving patterns. Ants use trail to mark paths to be followed by other nestmates [7], including species-specific products from various glandular sources (Dufour’s gland, hindgut, and poison gland) with varying composition and effectiveness [8]. Worker ants follow trail pheromones in a weaving (“zig-zag”) pattern [9]. Although ants may use other mechanisms to find a food source [10], ants detect trail pheromones via receptors in their antennae. As a worker ant follows a pheromone trail, the ant tends to move in and out of an active vapor space from one side then to the other. As one , containing the odor receptors, leaves the active vapor space the ant swings back in the opposite direction exhibiting a zig-zag pattern during its time on the pheromone trail [11]. How much a certain ant deviates from a straight line as it travels between two points determines how efficient the ant movement is. Despite this general pattern, differences in the parameters that define the weaving along an established

Insects 2020, 11, 5; doi:10.3390/insects11010005 www.mdpi.com/journal/insects Insects 2020, 11, 5 2 of 9 trail have not been recorded for many pest ants. The purpose of this laboratory study was to compare characteristics of foraging trails of pestiferous ants, both leaving their harborage toward a food source and returning from a food source back to their nest. Trail-following parameters may be important in understanding the efficiency of ant trail-following movement. The understanding of normal trailing following parameters but as the ant moves away from the nest, and as it moves toward the nest, it can provide important insight on the efficiency of different ant species as they forage for food. Comparison of the trailing efficiency between the paths away from the nest and the return to the nest provides that may be useful in testing new ant control products. Also, this information is useful in testing specific strategies and products that may disrupt trail-following or decrease the efficiency of the ant trail-following movement.

2. Materials and Methods Trail following patterns were compared for Nylanderia pubens (Forel) and Pheidole megacephala (Fabricius), Solenopsis invicta (Buren), Paratrechina longicornis (Forel), Technomyrmex albipes (Smith) and L. humile. Laboratory colonies were established from N. fulva, S. invicta, P. longicornis and T. albipes field collections at various sites around Gainesville, FL (Alachua County, FL, USA). Colonies of P. megacephala were collected in Gainesville, Dade City (Pasco County, FL, USA) and Inverness (Citrus County, FL, USA), and L. humile were collected from Lake City (Columbia County, FL, USA), in Florida. All laboratory ant colonies contained queens, brood and approximately 2000–4000 workers which were maintained at room temperature (22–27 ◦C), 64–70% RH and a photoperiod of 12:12 (L:D) h. Colonies were reared in 56 by 42 by 13 cm plastic trays with the inside walls coated with Fluon® (-a-slip, Bioquip Products Inc., Rancho Dominquez, CA, USA) to prevent ants from escaping. Ants were fed a regular diet of 10% sugar water, deionized water, boiled eggs, whole crickets, , house flies, grapes and orange slices. Nest cells (150 by 15 mm Petri dish) with a plaster (Castone™ dental stone) bottom, moistened regularly, were provided for to use as harborage. Foraging arenas were clear plastic boxes (27 by 19 by 9.5 cm, Pioneer Plastics, Dixon, KY, USA) with inside walls lined with Fluon®. Each arena contained 400 workers, 0.10 g of brood and 1 or 2 queens per tray. Queens and brood were included in the arenas to encourage worker foraging and trail following. Previous pilot studies confirmed that queens and brood presence were necessary stimuli for the workers to actively forage and seek nutrient resources. A small nest cell (100 by 15 mm Petri dish) with a moistened plaster bottom was provided for harborage. Ants were provided no food for 24–72 h prior to trials with access to water in plastic vials (86 by 33 mm, #15 dram vials, Thornton Plastics Co, Salt Lake City, UT, USA) with a cotton stopper. For trials involving T. albipes, an additional Pyrex® test tube (150 by 16 mm, Corning Inc., Corning, NY, USA) with a cotton stopper containing 10% sugar water was added to prevent high mortality during the starvation period. During the acclimation period, all worker ants and queens nested inside the small Petri dishes prior to experiments. Trials were conducted in a complete randomized design (CRD) with ant species as treatments and trials serving as experimental replication. Five trials were conducted with each of the six ant species. A deli cup (242 mL, American Plastics, Chattanooga, TN, USA) with the inside and outside walls lined with Fluon® was placed in the foraging arenas to serve as the site of food placement. A small slice of hard-boiled egg (~0.1 g), a slice of orange (~0.3 g) and one whole cricket (~0.8 g) were placed inside the deli cup. Ants could only access the food inside the deli cup via a wooden bridge made of red oak (20 by 4 cm by 6.5 cm high) (Figure1). On top of the wooden bridges, a study section (15 by 4 cm) was marked with a graphite #2 pencil consisting of a 5 mm 5 mm grid of marks, with longitudinal lines × throughout the entire wooden bridge. The grid was used to determine actual position of each ant on the wooden bridge at any time. The start and endpoints of the study area were also marked on top of the wooden bridge. The bridge represented an obstacle-free, simple flat surface so that ant trail-following behavior could be studied with minimum influence from the environment, and so all ant species would be exposed to the same environment to minimize any outside influences on trail-following. Insects 2019, 10, x FOR PEER REVIEW 3 of 9 and so all ant species would be exposed to the same environment to minimize any outside influences on trail-following. Trail pheromone extracts from each ant species were obtained by crushing the gasters of 30 workers with a glass mallet and washing in 2 mL of hexane (C6H14, Fischer Chemicals, Fair Lawn, NJ, USA). A disposable pipette was used to transfer the hexane/pheromone solution into scintillation vials. The extracts were prepared fresh when needed and refrigerated (4 °C) in a capped scintillation vial for at least 1 h prior to use. To encourage foraging in the test arenas a trail was marked on the bridge with a trail pheromone for each individual ant species. The trail pheromone was applied to the center of the wooden bridge (piles and deck) with a calligraphy pen (Koh-I-Noor® Rapidograph, Ceske Budejovice, Czech Republic) with a tip size number 0.0 (liquid line width of 0.30 mm) for a total of 0.20 mL (0.09 ant worker equivalent/cm) of trail pheromone applied on the wooden bridges. Separate bridges were used for each ant species, and any new trial, to avoid trail pheromone contamination. A digital video camera (Sony Handycam model DCR-HC36, Sony Electronics Inc., San Diego, CA, USA) was used to record the ant trail-following patterns and a microcomputer software program for time-lapse video editing (Adobe Premier Elements 4.0, 2008) was used to process the videos and make necessary measurements. Experiments were recorded for 30 min with the digital video camera Insectsmounted2020, 6111, 5cm above. Ant trail-following paths (going to and returning from food source) in3 the of 9 videos were transposed onto graph paper.

Figure 1. WoodenWooden bridges (40 × 150150 mm mm with with 65 65 mm mm elevatio elevationn from from arena arena floor) floor) used used in in trail-following trail-following × arenas. Black lines and dotes were spaced in a 5 × 55 mm mm grid grid and served as guides for measurements × of ant movement in the study area, ant path deviationsdeviations away from the longitudinal pheromone trail (amplitude), andand quantification quantification of theof antthe movement.ant movement. A gray A longitudinal gray longitudinal band was band added was graphically added tographically the picture to tothe represent picture to the represen centrallyt the applied centrally trail applied pheromone. trail pheromone.

TrailIn order pheromone to quantify extracts the from experimental each ant species parameters, were obtained the byfollowing crushing thecriteria gasters were of 30 used workers in withestablishing a glass mallethow ants and were washing to be in counted 2 mL of hexaneto avoid (C6H14, biases: Fischer(a) only Chemicals, ants starting Fair the Lawn, trail NJ,within USA). a Amaximum disposable of 5 pipette mm deviation was used from to transferthe center the of hexanethe bridge/pheromone (pheromone solution trail) intowere scintillation counted; (b) vials.ants Thewere extracts counted were for 10 prepared min maximum, fresh when or until needed 15 andants refrigeratedwere counted, (4 ◦ whicheverC) in a capped came scintillation first; (c) the vial initial for at10 leastmin of 1 hvideotape prior to use. footage To encourage was not used foraging in order in the totest allow arenas the scout a trail ants was to marked find the on pheromone the bridge withtrail aand trail recruit pheromone workers for during each individual this initial ant time species. period. The For trail all pheromoneants that started was applied foraging to down the center the center of the woodentrail in either bridge direction, (piles and during deck) the with 10-min a calligraphy video footage pen (Koh-I-Noor examined ®a Rapidograph,proportion of Ceskeants that Budejovice, finished Czechthe trail Republic) was calculated. with a tip size number 0.0 (liquid line width of 0.30 mm) for a total of 0.20 mL (0.09 ant worker equivalent/cm) of trail pheromone applied on the wooden bridges. Separate bridges were used for each ant species, and any new trial, to avoid trail pheromone contamination.

A digital video camera (Sony Handycam model DCR-HC36, Sony Electronics Inc., San Diego, CA, USA) was used to record the ant trail-following patterns and a microcomputer software program for time-lapse video editing (Adobe Premier Elements 4.0, 2008) was used to process the videos and make necessary measurements. Experiments were recorded for 30 min with the digital video camera mounted 61 cm above. Ant trail-following paths (going to and returning from food source) in the videos were transposed onto graph paper. In order to quantify the experimental parameters, the following criteria were used in establishing how ants were to be counted to avoid biases: (a) only ants starting the trail within a maximum of 5 mm deviation from the center of the bridge (pheromone trail) were counted; (b) ants were counted for 10 min maximum, or until 15 ants were counted, whichever came first; (c) the initial 10 min of videotape footage was not used in order to allow the scout ants to find the pheromone trail and recruit workers during this initial time period. For all ants that started foraging down the center trail in either direction, during the 10-min video footage examined a proportion of ants that finished the trail was calculated. The trail-following behavior of a particular ant species was assessed by quantifying trail-following ability recorded on graph paper into the following six parameters: (a) velocity (b) amplitude, (c) fidelity, (c) number of turns, (d) trail-following accuracy, and (e) proportion of ants finishing trails. Amplitude Insects 2020, 11, 5 4 of 9 is the measurement (in mm) of the perpendicular distance each ant species deviates from the center trail pheromone. Fidelity represents the percentage of the total bridge distance that the ant travels on top of the pheromone trail (within 1 mm of the marked trail). Turns (total number of) measures the frequency of corrective actions (left and right turns away or towards the center trail) an ant takes as it follows the trail. Velocity (mm/s) is calculated by dividing the time the ant takes to complete the 150 mm marked trail by the traveled distance. Finishing proportion measures the ratio of ants that successfully complete the trail (going to/returning from a food source) during the 10-min interval. Trail-fallowing accuracy was calculated by the following formula:   s    " #!2    (d f )   TA = d 2t − + a2 + f  100,      ÷  ∗ 2t   × where TA = percent trail-following accuracy, d = trail distance (mm), t = number of turns f = fidelity (mm), and a = amplitude (mm). This formula considers that turns in the ant path increase the distance an ant has to walk between two points on the trail, and that the increase is related to the number of curves and how far the ant deviates from a straight path. The formula calculates the ant path as a series of straight lines either on the direct pheromone path or from the direct path to a point that is 1 amplitude unit away from the straight path and then back to the straight path. The length of the path that takes the ant away from, and then back to the marked pheromone trail is calculated as the hypotenuse of the right-angled triangle that has as its basis the straight line on the pheromone path, and the amplitude of the path as the height of the triangle. Given that the fidelity is the length of the path the ant stays right on the trail pheromone, the total trail distance minus the fidelity represents the portion of the ant path that deviates from the pheromone-marked trail. Thus, the total path followed by the ant can be mathematically calculated as a series of straight lines on the pheromone trials plus right-angle triangular deviations from and back to the pheromone. Since all ants were exposed to the same exact terrain on the top of the bridge, and comparisons were only made for trail-following in this uniform portion of the ant paths, direct comparison of the trails from different ants was possible. Trail-following accuracy, as presented here, is a measure of how efficient an ant species is when traveling on a straight path. Trail-following accuracy is not meant to measure the ability of an ant species to detect and follow a pheromone trail as proposed by other authors [12–14]. Statistical analyses were performed in JMP 13.2.0 (SAS Institute 2016). One-way ANOVA’s were used to evaluate differences in the population means with ant species as the main factor and the five parameters: amplitude, fidelity, turns, trail-following accuracy and velocity as the response variables. Treatment means were separated by Student’s t-test. The level of significance for all statistical analysis was set at α = 0.05.

3. Results The mean velocity for the different ants varied from a maximum of 28.1 0.82 mm/s for P.longicornis ± to a minimum of 9.6 0.16 mm/s for S. invicta (Figure2). There was no significant di fference in the ± mean velocity between P. megacephala (17.0 0.39 mm/s) and L. humile (17.0 0.24 mm/s), but all other ± ± ant species moved at a different mean velocity. Paratrechina longicornis and P. megacephala workers moved faster going toward food (29.5 1.35 and 18.5 0.53 mm/s, respectively) than returning to the ± ± nest (26.8 0.90 mm/s and 15.5 0.54 mm/s, respectively). For the other ants (N. fulva, 23.5 0.38; ± ± ± L. humile, 17.0 0.24; T. albipes, 13.6 0.27; and S. invicta, 9.6 0.16 mm/s) there were no significant ± ± ± differences between the velocity of ant trail-following to and from the food source. InsectsInsects 20202019,, 1110,, 5x FOR PEER REVIEW 55 ofof 99 Insects 2019, 10, x FOR PEER REVIEW 5 of 9

FigureFigure 2.2. Mean ant velocity (mm(mm/s)/s) (mean(mean ± 95% CI) as ants move towards a food source (Foraging) Figure 2. Mean ant velocity (mm/s) (mean ±± 95% CI) as ants move towards a food source (Foraging) andand returningreturning toto thethe nestnest (Returning).(Returning). Means toppe toppedd by didifferentfferent letters are significantlysignificantly didifferentfferent and returning to the nest (Returning). Means topped by different letters are significantly different (Students(Students meansmeans separation;separation; αα == 0.05). (Students means separation; α = 0.05). ThereThere waswas no no significant significant diff differenceerence in mean in mean path amplitudepath amplitude between between the foraging the foraging and the returning and the pathsreturningThere for any paths was ant, nofor except significantany for ant,P. megacephalaexcept difference for (FigureP. in megacephala mean3), whichpath (Figure amplitude deviated 3), more whichbetween from deviated thethe marked foraging more trail from and as thethe antsmarkedreturning returned trail paths as to the thefor ants nestany returned(1.7 ant, except0.11 to mm) the for nest than P. megacephala(1.7 when ± 0.11 ants mm) went(Figure than out 3), foragingwhen which ants (1.1 deviated went0.03 out mm).more foraging However,from (1.1 the ± ± ± there0.03marked mm). were trail However, significant as the ants there di returnedff erenceswere significant to among the nest thedifference (1.7 di ±ff erent0.11s mm)among ant than species the when different in relationants antwent species to out path foraging in amplitudes relation (1.1 to ± (path0.03F = 82.2 mm).amplitudes, p However,< 0.0001), (F = with82.2,there P.p were < longicornis 0.0001), significant with(3.0 P. difference 0.10longicornis mm)s andamong (3.0N. ± 0.10 pubens the mm)different(2.2 and0.08 N.ant pubens mm)species showing (2.2 in ±relation 0.08 greater mm) to ± ± pathshowingpath amplitudesamplitudes greater than (Fpath = T.82.2, albipesamplitudes p < 0.0001),(1.7 0.07than with mm), T. P. albipes longicornisS. invicta (1.7(1.5 (3.0± 0.07 ±0.08 0.10 mm), mm), mm) S.P. and megacephalainvicta N. pubens (1.5 (± 1.4(2.2 0.08 ±0.06 0.08mm), mmmm) P.), ± ± ± andmegacephalashowingL. humile greater (1.4(1.3 ± path0.060.03 mm),amplitudes mm). and L. humilethan T. (1.3 albipes ± 0.03 (1.7 mm). ± 0.07 mm), S. invicta (1.5 ± 0.08 mm), P. megacephala (1.4 ± ±0.06 mm), and L. humile (1.3 ± 0.03 mm).

Figure 3. Mean trail amplitude (mm) (mean 95% CI) as ants move towards a food source (foraging) Figure 3. Mean trail amplitude (mm) (mean ± 95% CI) as ants move towards a food source (foraging) and returning to the nest (returning). Ant species topped by different letters are significantly different Figure 3. Mean trail amplitude (mm) (mean ± 95% CI) as ants move towards a food source (foraging) (studentsand returning means to separation;the nest (returning).α = 0.05) Ant There species were to nopped significant by different differences letters betweenare significantly the foraging different and and returning to the nest (returning). Ant species topped by different letters are significantly different returning(students pathsmeans for separation; any ants exceptα = 0.05) for ThereP. megacephala were no significant(marked with differences *). between the foraging and (studentsreturning meanspaths for separation; any ants αexcept = 0.05) for There P. megacephala were no significant (marked with differences *). between the foraging and Thereturning mean paths number for any of turnsants except individual for P. megacephala ants took towards(marked with and *). away from the center pheromone trail wereThe mean significantly number di offf erentturns (Figureindividual4; F =ants82.8; took df =towards5; p = < and0.0001) away among from N.the pubens center(15.1 pheromone0.28), ± L.trail humile wereThe (1.5mean significantly 0.18),numberT. albipesdifferent of turns(10.3 individual(Figure0.18); 4; andF ants = 82.8;P. took longicornis df towards = 5; p (9.69= <0.0001)and away0.15), among from but there N.the pubens center was no (15.1 pheromone significant ± 0.28), ± ± ± ditrailL.ff humileerence were (1.5 insignificantly the ± 0.18), mean T. number albipesdifferent of(10.3 turns(Figure ± 0.18); between 4; andF = 82.8; S.P. invictalongicornis df = (13.25; p =(9.69 <0.0001)0.25) ± 0.15), and among big-headed but there N. pubens was ants no ((15.112.8 significant ± 0.28),0.25), differenceL. humile (1.5 in ±the 0.18), mean T. albipesnumber (10.3 of turns± 0.18); between and P. longicornisS. invicta (13.2(9.69± ±± 0.25)0.15), andbut therebig-headed was no ants significant± (12.8 ± difference in the mean number of turns between S. invicta (13.2 ± 0.25) and big-headed ants (12.8 ±

Insects 20192020, 1011, x 5 FOR PEER REVIEW 6 of 9 Insects 2019, 10, x FOR PEER REVIEW 6 of 9 0.25), although these two ants were also significantly different from all other ants. All ants had a similaralthough0.25), although number these twotheseof turns ants two wereboth ants alsogoingwere significantly alsoto and significantl coming different yfrom different from food, all from except other all ants. forother S. All ants.invicta ants All hadwhich ants a similar madehad a significantlynumbersimilar number of turns more of both turns turns going going both to toward andgoing coming tothe and food from coming (14.1 food, ± from0.37) except thanfood, for going S.except invicta back forwhich to S. the invicta made nest (12.2which significantly ± 0.32).made moresignificantly turns going more toward turns going the food toward (14.1 the0.37) food than(14.1 going ± 0.37) back than to going the nest back (12.2 to the0.32). nest (12.2 ± 0.32). ± ±

Figure 4. Mean total number of turns (mean ± sem) as ants move towards a food source (foraging) Figure 4. Mean total number of turns (mean sem) as ants move towards a food source (foraging) andFigure returning 4. Mean to totalthe nest number (returning). of turns Ant (mean species ± sem) topped as antsby different move towards letters are a food significantly source (foraging) different and returning to the nest (returning). Ant species topped by different letters are significantly different (studentsand returning means to separation;the nest (returning). α = 0.05). Ant There species were tonopped significant by different differences letters between are significantly the foraging different and (students means separation; α = 0.05). There were no significant differences between the foraging and returning(students meanspaths within separation; the sa αme = 0.05).ant species, There wereexcept no for significant S. invicta differences (marked with between *). the foraging and returning paths within the same ant species, except for S. invicta (marked with *). returning paths within the same ant species, except for S. invicta (marked with *). The mean percent fidelity fidelity did not vary between the paths going toward the food and going back The mean percent fidelity did not vary between the paths going toward the food and going back to the nest for P. longicornislongicornis (16.4(16.4 ± 0.62), N. fulva (30.7 ± 0.81)0.81),, and L. humile (36.1 ± 0.82)0.82) (Figure (Figure 55).). to the nest for P. longicornis (16.4 ±± 0.62), N. fulva (30.7 ±± 0.81), and L. humile (36.1 ± 0.82) (Figure 5). However, both P. megacephala and T. albipesalbipes had greater trail fidelityfidelity going toward the food than However, both P. megacephala and T. albipes had greater trail fidelity going toward the food than returning to the nest, while S. invicta invicta hadhad the the opposite opposite behavior, behavior, having greater trail fidelity fidelity returning returning to the nest, while S. invicta had the opposite behavior, having greater trail fidelity returning to the nest. to the nest.

Figure 5. Mean percent trail fidelity (mean sem) as ants move towards a food source (foraging) and Figure 5. Mean percent trail fidelity (mean ± sem) as ants move towards a food source (foraging) and returningFigure 5. Meanto the percentnest nest (returning). (returning). trail fidelity Means Means (mean topped topped ± sem) by by dias different ffantserent move letters letters towards are are signific significantly a foodantly source different diff erent(foraging) (students and α meansreturning separation; to the nest α (returning).= 0.05). Significant Significant Means topped differences differences by di betweenfferent between letters the the foraging are signific and antly returning different paths (students were detectedmeans separation; for P. megacephala,megacephala, α = 0.05). T.T. Significant albipesalbipes,, andand differences S.S. invictainvicta.. between the foraging and returning paths were detected for P. megacephala, T. albipes, and S. invicta.

Insects 2020, 11, 5 7 of 9 Insects 2019, 10, x FOR PEER REVIEW 7 of 9

There asas a a significant significant di ffdifferenceerence (F = (F68.2, = 68.2,p < 0.0001) p < 0.0001) in the trail-followingin the trail-following accuracy accuracy which basically which putbasically the ants put in the two ants groups, in two with groups, the crazy with antsthe crazy having ants lower having trail-following lower trail-following accuracy (86–92%) accuracy and (86%– the other92%) antsand withthe other higher ants accuracy with (94–98%)higher accuracy (Figure6 (94%–98%)). There were (Figure no significant 6). There di ffwereerences no insignificant accuracy (Fdifferences= 3.19, p =in0.07) accuracy between (F = the3.19, track p = 0.07) toward between the food the and track the toward track back the food to the and nest the for track any back of the to ants, the althoughnest for any the of general the ants, trend although for most the of general the ants tren wasd afor slightly most lowerof the trail-followingants was a slightly accuracy lower for trail- the returnfollowing track. accuracy However, for wethe detectedreturn track. no di ffHowever,erences in we the detected proportion no ofdifferences ants finishing in the trails, proportion for any of theants tested finishing species, trails, either for any going of the to ortested returning species, from either the food.going to or returning from the food.

Figure 6. Mean trail-following accuracy (mean sem) as ants move towards a food source (Foraging) Figure 6. Mean trail-following accuracy (mean ± sem) as ants move towards a food source (Foraging) and returning to the nest (Returning). Ant species topped by different letters are significantly different and returning to the nest (Returning). Ant species topped by different letters are significantly different (Students’s means separation; α = 0.05). There were no significant differences between the foraging (Students’s means separation; α = 0.05). There were no significant differences between the foraging and returning paths within the same ant species. and returning paths within the same ant species. Despite the different trail parameters being used in the calculation of the trail-following accuracy, the onlyDespite parameter the different with high trail correlation parameters (r 2 being= 0.400–0.751) used in withthe calculation the trail-following of the trail-following accuracy was 2 theaccuracy, amplitude. the only Greater parameter amplitude with high led tocorrelation lower trail-following (r = 0.400–0.751) accuracy, with the whereas trail-following lower amplitude accuracy correspondedwas the amplitude. well with Greater higher amplitude trail-following led accuracy.to lower Notrail-following statistically significantaccuracy, correlationwhereas lower was observedamplitude between corresponded the other well trail with parameters higher measured trail-following in this study.accuracy. No statistically significant correlationThere waswas noobserved significant between difference the other among trail the parameters ant species measured for the proportion in this study. of starting ants that actuallyThere completed was no significant the trail, (F difference= 0.29, p = among0.92) or the between ant species ants for going the awayproportion from the of starting nest of returning ants that toactually the nest completed (F = 0.44, thep = trail,0.51), (F with= 0.29, an p 81.8= 0.92)10.41% or between completion ants going rate away for ants from moving the nest away of returning from the ± nestto the and nest a 79.6(F = 0.44,11.94% p = 0.51), completion with an rate 81.8 for ± ants10.41% coming completion back to rate the nest.for ants moving away from the nest and a 79.6 ± 11.94% completion rate for ants coming back to the nest. 4. Discussions 4. Discussions Clear behavioral differences in trail-following patterns are demonstrated here for six major pest ant species,Clear behavioral with an obvious differences separation in trail-following between the crazypatterns ants, areP. demonstrated longicornis and hereN. fulva for ,six and major the other pest pestant species, ant species with in an several obvious aspects separation of their between trail-following the crazy behavior. ants, P. The longicornis most striking and N. di fffulvaerence, and in the measuredother pest parametersant species in is onseveral the trail-following aspects of their speed, trail-following with the crazybehavior. ants showingThe most astriking 2 to 3 times difference faster pacein the than measured other ants, parameters but also is with on the greater trail-followin amplitudeg speed, and lower with trail the fidelitycrazy ants than showing the other a 2 tested to 3 times ants, resultingfaster pace in than a significantly other ants, lower but also trail-following with greate accuracyr amplitude for the and crazy lower ants. trail A fidelity behavioral than adaptation the other totested moving ants, fairlyresulting quickly in a significantly and more erratically lower trail-following outside of the accuracy nest has for the the advantage crazy ants. of A preventingbehavioral exposureadaptation to to both moving the elements fairly quickly and predation. and more Conversely, erratically theoutside very aggressiveof the nest firehas ant the had advantage the lowest of trailingpreventing velocity, exposure which to mayboth bethe related elements to itsand high pred degreeation. ofConversely, aggressivity the [ 15very]. aggressive fire ant had the lowestPotentially, trailing it velocity, can be advantageous which may be inrelated maximizing to its high food degree discovery. of aggressivity However, [15]. despite all the differencesPotentially, observed it can in be the advantageous trail parameters in ofmaximi the studyzing ants,food thediscovery. ability of However, the ants todespite complete all the traildifferences was fairly observed consistent in the among trail parameters the different of species. the study This ants, indicates the ability that theof the diff antserent to trail-following complete the trail was fairly consistent among the different species. This indicates that the different trail-following

Insects 2020, 11, 5 8 of 9 strategies may be more a result of physical characteristics of the ants and less a consequence of an evolutionary process affecting the trail-following behavior. There were some interesting discrepancies between the different parameters measured or calculated for the foraging behavior of the study ants. For instance, despite N. fulva staying on the marked trail approximately twice as long as P. longicornis, trail-following accuracy for the tawny crazy ants was significantly lower than that for the black crazy ant. The tawny crazy ant seems to follow the marked trail fairly well, but when it deviates from the marked trail, those deviations either take the ants farther from the marked trails or for longer periods. Comparatively, the other ants in the study had fairly high fidelity to the marked trail, which, combined with a low amplitude of the deviations, resulted in high trail-following accuracy for P. megacephala, L. humile, T. albipes and S. invicta. Because ant trails are normally bidirectional [16], and, in our experiments, ants were encouraged to maintain that aspect because of a single marked path on the wooden bridge, some of the movements away from the marked path were dictated by the presence of other ants traveling in the opposite direction. Differences in the fidelity between ants leaving the nest and ants returning to the nest are a consequence of this bidirectional movement. The ants’ fidelity to the trail fall into three different patterns in the study ants: (a) ants with similar fidelity whether moving away or toward the nest, represented by P. longicornis, N. fulva, and L. humile; b) ants with greater fidelity to the trail when moving away from the nest, represented by P. megacephala and T. albipes; and (c) ants with greater fidelity on the return trip to the nest, represented by S. invicta. Greater fidelity to the trail on the return trip to the nest would seem to be the preferred behavior, assuming that an ant laden with food should move more directly to the nest, both the minimize risks and to save energy. This is especially true when ants are exposed to risks present in the field that were absent in our studies. However, low trail fidelity can be advantageous to ants [17]. Despite our measurements being taken in a very uniform wooden path with a well-marked path for the ants to follow, trail-following accuracy showed great variation between the group of the crazy ants with trail-following accuracy at 91% and below, and the other ants in the study, which despite considerable taxonomic and ecological differences, had fairly uniform trail-following accuracy between 95% and 97%. Trail accuracy seems to be the best measurement of the erratic behavior of the crazy ants, which can be detected even when the ants are following precise trails and not responding to outside stimulus as is the case when ants are disturbed during foraging. In this work, the trail-following patterns are quantitatively described for the six major pest ant species using parameters that quantified the zig-zag trail-following pattern in ants. Understanding trail-following behaviors of invasive pest ants is an important factor that may allow a greater understanding of why these ants become efficient pestiferous species. Also, a better understanding of foraging behaviors of pest ants may play a role in the development and refinement of control strategies that best suit different pest ant species, such as toxic bait placement around the established trail, and application of products that may disrupt trail-following behavior.

5. Conclusions Trail-following patterns for major pest ant species vary in relation to velocity, amplitude, fidelity, number of turns, and trail-following accuracy, falling into three different patterns with (a) similar trail fidelity whether moving away or toward the nest; (b) greater fidelity to the trail when moving away from the nest; and (c) greater fidelity on the return trip to the nest. Trail-following behavior is an important factor in understanding the efficiency of pest ant species.

Author Contributions: Conceptualization, R.J.V. and R.M.P.; methodology, R.J.V. and R.M.P.; formal analysis, R.J.V. and R.M.P.; resources, P.G.K.; data curation, R.J.V.; writing—original draft preparation, R.J.V.; writing—review and editing, R.M.P. and P.G.K.; supervision, P.G.K.; project administration, P.G.K.; funding acquisition, P.G.K. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Insects 2020, 11, 5 9 of 9

Acknowledgments: Assistance with ant rearing provided E. Pereira. Video recording and experiment assistance by. M. Mitola. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Halley, J.D.; Burd, M. Nonequilibrium dynamics of social groups: Insights from foraging Argentine ants. Insectes Sociaux 2004, 51, 226–231. [CrossRef] 2. Meissner, H.E.; Silverman, J. Effect of aromatic cedar mulch on Argentine ant (Hymenoptera: Formicidae) foraging activity and nest establishment. J. Econ. Entomol. 2003, 96, 850–855. [CrossRef][PubMed] 3. Vega, S.Y.; Rust, M.K. Determining the foraging range and origin of resurgence after treatment of Argentine ant (Hymenoptera: Formicidae) in urban areas. J. Econ. Entomol. 2003, 96, 844–849. [CrossRef][PubMed] 4. Cole, B.J.; Edwards, R.; Holbrook, C.T.; Holm, L.; Heyward, J.; Wiernasz, D.C. Does foraging activity affect foraging success in the western harvester ant (Hymenoptera: Formicidae)? Ann. Entomol. Soc. Am. 2008, 101, 272–276. [CrossRef] 5. Wetterer, J.K.; Gruner, D.S.; Lopez, J.E. Foraging and nesting ecology of Acromyrmex octospinosus (Hymenoptera: Formicidae) in a Costa Rican tropical dry forest. Fla. Entomol. 1998, 81, 61–67. [CrossRef] 6. Traniello, J.F.A. Foraging strategies of ants. Annu. Rev. Entomol. 1989, 34, 191–210. [CrossRef] 7. Czaczkes, T.J.; Grüter, C.; Jones, S.M.; Francis, L.W.; Ratnieks, F.L.W. Uncovering the complexity of ant foraging trails. Commun. Integr. Biol. 2012, 5, 78–80. [CrossRef][PubMed] 8. Wilson, E.O. The Insect Societies; Harvard University Press: Cambridge, MA, USA, 1971; p. 548. 9. Hangartner, W. Spezifitat und Inaktivierung des Spurpheromons von Lasius fuliginosus Latr. und Orientierung der Arbeiterinnen im Duftfeld. Z. Vergl. Physiol. 1967, 57, 103–106. [CrossRef] 10. Czaczkes, T.J.; Grüter, C.; Jones, S.M.; Francis, L.W.; Ratnieks, F.L.W. Synergy between social and private information increases foraging efficiency in ants. Biol. Lett. 2011, 7, 521–524. [CrossRef][PubMed] 11. Holldobler, B.D.; Wilson, E.O. The Ants; Belknap Press: Cambridge, MA, USA, 1990. 12. Czaczkes, T.J.; Castorena, M.; Schürch, R.; Heinze, J. Pheromone trail following in the ant Lasius niger: High accuracy and variability but no effect of task state. Physiol. Entomol. 2017, 42, 91–97. [CrossRef] 13. von Thienen, W.; Metzler, D.; Choe, D.-H.; Witte, V. Pheromone communication in ants: A detailed analysis of concentration-dependent decisions in three species. Behav. Ecol. Sociobiol. 2014, 68, 1611–1627. [CrossRef] 14. Grüter, C.; Czaczkes, T.J.; Ratnieks, F.L.W. Decision making in ant foragers (Lasius niger) facing conflicting private and social information. Behav Ecol Sociobiol. 2011, 65, 141–148. [CrossRef] 15. Slipi´nski,P.;´ Zmihorski,˙ M. Changes in the speed of ants as a result of aggressive interactions. Insect Sci. 2017, 24, 842–852. [CrossRef][PubMed] 16. Fourcassié, V.; Dussutour, A.; Deneubourg, J.-L. Ant traffic rules. J. Exp. Biol. 2010, 213, 2357–2363. [CrossRef] [PubMed] 17. Elizalde, L.; Farji-Brener, A. To be or not to be faithful: Flexible fidelity to foraging trails in the leaf-cutting ant Acromyrmex lobicornis. Ecol. Entomol. 2012, 37, 370–376. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).