Brood-Holding Causes Workers to Pay Attention to the Queen in The

bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 [Title] 2 3 Brood-holding causes workers to pay attention to the queen in the carpenter 4 ant Camponotus japonicus 5 6 7 Author’s name: 8 Kenji Hara 9 10 Address: 11 Life Science, Tokyo Gakugei University, 4-1-1, Nukui-Kitamachi, 12 Koganei-shi, Tokyo 184-8501, Japan. 13 14 Abbreviated form of the title: 15 Attention of the ant workers to the queen 16 17 Name and address for correspondence: 18 Kenji Hara (Ph.D.) 19 Life Science, Tokyo Gakugei University, 4-1-1, Nukui-Kitamachi, 20 Koganei-shi, Tokyo 184-8501, Japan. 21 Tel/FAX; +81-42-329-7522 22 e-mail; [email protected] 23

1 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 [Abstract] 2 3 Brood accumulation, a fundamental behavior of offspring care in the carpenter ant 4 Camponotus japonicus, is driven by alternation of 'holding run' and 'empty-handed 5 run' behaviors. In the holding run, a worker holds a brood with her mandibles and 6 carries it to the queen (holding run). After releasing it beside the queen, she hurries 7 back to another brood (empty-handed run). To address the motivation for the 8 brood-accumulation task, in this study, I observed these behaviors under 9 experimental conditions. When workers performed the task in a situation that 10 involved selection between their own and unfamiliar queens, they ran in 11 significantly more restrictive ways during the holding run than during the 12 empty-handed run. Hence, 'holding' represents a different motivational state than 13 'empty-handed'. In a second experiment, the workers were suddenly presented with 14 an unfamiliar floor during the task. Regardless of whether they were holding or 15 empty-handed, their running traces on the familiar floor were simple, whereas on 16 the unfamiliar floor they were more complex. These results show that holding 17 workers would pay attention to the queen, exploiting cues on the floor to restrict 18 their responses to the queen. 19 20 [Key words] 21 22 selective attention, olfaction, reiteration, social behavior, insect 23

2 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 [Introduction] 2 3 The behavioral complexity of insects has been demonstrated in behavioral 4 and physiological studies of many species. Insects' adaptive actions are based on 5 reflexes and/or the internal patterns, modulated by their response to sensory 6 stimuli. Motivation is a process that leads to the formation of behavioral intentions, 7 i.e., it sets the aim of the behavior. In a behavior that is accomplished by combining 8 multiple actions together, there is thought to be a different motivation for every 9 action. For example, desert ants, Cataglyphis fortis, use path integration (PI) as 10 their main mode of navigation (Wehner, 2003). Their foragers continuously measure 11 directions and distances both when going out to a feeding site (outbound) and when 12 returning to the nest (inbound), and then form outbound and inbound vectors, 13 respectively, by integrating these two quantities. Utilization and switching of the 14 vectors associated with the motivation allows the ants to perform the adaptive 15 behavior (Collett et al., 1999; Merkle and Wehner, 2008). The appropriate change in 16 motivation is necessary to achieve the goal resulting from a chain of actions. 17 In social insects, brood care by sterile workers plays important roles in 18 maintenance of highly sophisticated communities. In the carpenter ant Camponotus 19 japonicus, nestmates divide labor according to their ages, and brood care is the 20 initial task after emergence. Nurse workers repeatedly carry scattered broods and 21 accumulate them beside the queen (Hara, 2013), who affects brood growth by 22 producing and secreting various substances (Holman, 2010; Motais de Narbonne et 23 al., 2016). Therefore, brood-accumulation behavior is fundamental to the altruistic 24 behavior responsible for rearing of kin broods by sterile workers. 25 The unit of brood-accumulation behavior consists of four sequential 26 elements: (1) The worker recognizes a brood by touching it with the antennae, and 27 then picks it up with her mandibles ('pickup'). (2) The worker starts an inbound run 28 while holding a brood ('holding'). (3) The worker recognizes the queen by touching 29 with the antennae, and then places the brood beside her ('release'). (4) The worker 30 turns her back on the queen and starts the outbound run ('empty-handed'). 31 Repetition of the unit results in gathering of broods near the queen. 32 To address the physiological mechanisms underlying brood accumulation behavior, 33 the workers performed tasks under two experimental conditions: a situation that 34 involved selection between unfamiliar (UQ) and fostered queens (FQ) (Experiment 35 1), and a situation in which the familiar floor disappeared (Experiment 2). The data 36 were analyzed to compare 'holding' and 'empty-handed' behaviors. I discuss the

3 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 results from the standpoint of attention-like phenomena of workers engaging in the 2 brood-accumulation task. 3 4

4 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 [Materials and methods] 2 3 Animals 4 5 Experiments were performed on the laboratory-reared workers and queens of 6 the carpenter ant Camponotus japonicus. The laboratory colonies were found and 7 maintained as previously described (Hara, 2002). Founding queens were collected 8 from 2004 to 2016 in Tokyo, Japan. In order to obtain the healthy data, the workers 9 and queens were used from the colonies within two years from the foundation. 10 To prepare the experimental workers, the pupae isolated from their birth 11 colonies were removed from the cocoon and were incubated individually in the 12 96-well culture plate with U bottom (Ishii et al., 2005). Shortly after emergence, 13 they were marked individually with cloth threads of different colors tied between 14 their petiole and gaster and were transferred to the foster colonies (Hara, 2003). On 15 acceptance by the natural ants and performance of the social activities, availability 16 of those workers was decided for this experiment. 17 18 Experimental design and trial procedure 19 20 Experiment 1: The brood-accumulation behaviors of the workers were 21 recorded under the selective condition between the foster queen (FQ) and an 22 unfamiliar queen (UQ). As details of the experimental procedure used for recording 23 the behavior were given in a previous paper (Hara, 2003), only a brief outline is 24 described below. An acrylic box was used for the test consisting of three rooms and 25 the central space (Fig. 1A). Each room is connected to the central space by a 26 doorway that allowed the workers free access to all rooms. A distance between the 27 corners of the central space (red dots in Fig. 1A) was 2.7cm. Two queens and broods 28 were put in the rooms of the box, Q1, Q2 and B, respectively. FQ was put in the 29 room Q1 or Q2, and UQ was put in the other room. After acclimation for 15 min, the 30 trial lasted 60 min. To avoid possible bias resulting from the UQ accidentally having 31 the 'colony labels' similar to those of FQ, the preliminary check were performed 32 every pair, with a control worker from the foster colony. 33 Experiment 2: To show an experimental worker the unfamiliar ground 34 condition suddenly on the trial, the following apparatus was used (Fig.1B); Two 35 plastic dishes (35 mm in diameter) were put in an acrylic box (252×345 mm). Broods 36 were set in the one designated as "B" in Fig. 1B and queen was in the other "Q". A

5 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 thread was attached to the dish "Q" in order to operate it from outside the box. FQ 2 was introduced in the apparatus and then a worker was allowed to explore the 3 inside for 15 min. After finishing such acclimation, five broods from the foster colony 4 were introduced into the dish "B" and then, the worker was put back into it. The 5 behavior had been recorded for 60 min by the video camera since the worker picked 6 a brood up. 7 The preliminary and the final tests were carried out sequentially with each 8 worker. The distance between the dishes "Q" and "B" (B-Q distance) was constant at 9 5 cm through the preliminary trial. When the worker released the 3rd brood beside 10 the queen in the final trial, the dish "Q" was moved away from the dish "B" to 10 cm 11 (Fig. 1B). In this study, the zone to 5cm at linear distance from "B" is referred to as 12 'familiar floor' and 'unfamiliar floor' showed the zone more outer than it. 13 14 Data analyses 15 16 All behaviors were recorded with the home video cameras (HDR-SR7, 17 HDR-CX180, SONY; HDC-TM90, Panasonic; GZ-MG330, Victor). All data were 18 captured at 30 frames per second. The video data were converted to AV1 format in 19 the computers. To quantify movement, the position of the ant's head was measured 20 frame-by-frame using the motion analysis software, DIPP-MortionPro2D (Ditect, 21 Japan). Two-dimensional coordinate data were saved as the excel file and were 22 taken advantage of for subsequent analysis. All statistical calculations were 23 performed with SPSS 22 and 23 (IBM, USA). 24

6 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 [Results] 2 3 Experiment 1; Effect of an unfamiliar queen as a distractor on brood-accumulation 4 behavior 5 6 1-1. Responses of workers to brood 7 In Experiment 1, I analyzed the responses of 149 laboratory-reared workers 8 over 12 years, from 2004 to 2015. The results are summarized in Table 1A. Within 3 9 days after emergence, most workers were largely immobile (standstill). Six workers 10 (21%) touched a brood with antennae, but five of them did not picked it up (i.e., they 11 ignored it). At 4–7 days of age, there was an increase in the number of workers who 12 held a brood: 13 (52%) held a brood and then released it beside another brood in the 13 same room (gathering), but never carried it out of room "B". The other five workers 14 (20%) picked up a brood and left room "B" while holding it (transportation). After 15 day 12, almost all workers performed transportation. 16 According to both repetition of the activity and the ability to discriminate 17 FQ from UQ, I performed further analysis of workers engaging in transportation; 18 the results are summarized in Table 1B. Of 91 workers, 77 repeatedly carried their 19 broods to the FQ. In this study, such brood-accumulation behavior characterized by 20 repetition and FQ recognition is described as 'regular'. Three workers brought some 21 or all of broods to the UQ. Eleven workers delivered only one brood to the FQ and, 22 after releasing it, did not leave the side of the FQ. No worker remained beside the 23 UQ after incorrect delivery. 24 25 1-2. Behaviors in the central space 26 For the 77 workers performing the regular task, I analyzed their trajectories 27 in the central space for each holding and empty-handed run. A typical example is 28 shown in Figure 2. During holding runs, the workers arrived at the FQ via 29 restricted courses (Fig. 2A, C). By contrast, during empty-handed runs, the workers 30 ran through a wider area and also seemed to be interested in the UQ (Fig. 2B, D). To 31 quantify the special activities of workers, I defined an array by consolidating the 32 position coordinates of the central space into a grid composed of 6 × 6 blocks; each 33 block was 0.45 cm × 0.45 cm square (lower panels in Fig. 2C and 2D). Array 34 analyses for 77 workers confirmed that the activity in the holding run was 35 significantly more restricted, both spatially (P<0.01, Wilcoxon test; Fig. 3A) and 36 temporally (P<0.01, test for the proportion; Fig. 3B) than in the empty-handed run.

7 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 2 1-3. Response of workers to queens 3 Queens induce behaviors in workers using pheromones. To determine 4 whether the queen’s signals would influence repeatability in the 5 brood-accumulation behavior, I analyzed the amounts of time each worker remained 6 with the FQ. At the first delivery, the workers (n=77) clung to the FQ, and 7 consequently stayed significantly longer in room "FQ" than during subsequent 8 deliveries (P<0.01, Wilcoxon test; Fig. 4A). In particular, after the third visit, the 9 workers left the room as soon as they released a brood beside the queen. 10 Few workers visited the UQ during the holding runs (Fig. 4B). Two workers 11 at both the first and second delivery stepped into room "UQ" and immediately 12 departed without touching the UQ. In the empty-handed runs, on the other hand, 13 the frequency of UQ visits was much higher (Fig. 4B): 60 of 77 workers (78%) 14 stepped into room "UQ" on the return trip after the first delivery. Some of them 15 touched the UQ with their antennae a few times, and consequently spent longer in 16 room "UQ". At the 2nd and subsequent deliveries, 28 (36%) and 10 (13%) of the 17 empty-handed workers, respectively, visited the UQ (Fig. 4B). 18 19 Experiment 2; Effect of an unfamiliar floor as a distractor on brood-accumulation 20 behavior 21 22 2-1. Responses of workers to brood 23 In Experiment 2, I analyzed responses to brood of 68 workers from 24 laboratory-reared colonies over 6 years from 2011 to 2016. The results are 25 summarized in Table 2A. In the preliminary test, 25 of 31 8–11-day-old workers and 26 36 of 37 12–14-day-old workers picked up a brood up and then carried it away from 27 dish "B" (transportation). For workers engaging in transportation in the 28 preliminary test, 21 of 25 8–11-day-old workers and all 36 12–14-day-old workers 29 also transported a brood out of dish "B" in the final test. Ultimately, 57 workers 30 (84%) had the ability to engage in brood transportation. 31 I performed further analysis of those 57 workers; the results are 32 summarized in Table 2B. 43 (75%) delivered five broods to the queen in both the 33 preliminary and final tests. Six workers (11%) repeatedly carried broods to the FQ, 34 but did not finish carrying all five broods in either or both tests (mission incomplete). 35 Three of those workers strayed in the empty-handed run just after extending the 36 B-Q distance and did not arrive at dish "B" before the time limit. Eight workers

8 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 (14%) delivered only the first brood to the FQ and, after releasing it, did not leave 2 the side of the FQ (no repetition). 3 4 2-2. Behaviors on the unfamiliar floor 5 This experiment sought to determine whether workers engaging in the 6 brood-accumulation task could continue their mission when they encountered a 7 ground condition. The workers became familiar with the floor space between the 8 broods and the queen by traversing it many times for brood delivery (Fig. 5A). In 9 the holding run for the third brood in the final test, they ran straight to the queen 10 (Fig. 5B). When the workers released the brood beside the queen, the B-Q distance 11 was expanded to present an unfamiliar floor. In the subsequent empty-handed runs, 12 the workers followed a meandering trace on the unfamiliar floor, whereas on the 13 familiar floor they ran straight and easily arrived at dish "B" (Fig. 5C, upper panel 14 in Fig. 5F). In the holding run for the fourth brood, the workers ran in an almost 15 constant direction on the familiar floor but began to stray once they stepped onto 16 the unfamiliar floor (Fig. 5D, upper panel in Fig. 5G). Their meandering runs were 17 maintained until the delivery was completed, even if they ran back onto the familiar 18 floor. 19 To quantify the effect of the ground condition on movement, I calculated 20 frame-by-frame the track angle of the running trajectory, defined by the running 21 vector and orientation to the goal; data were captured at 30 frames per second (Fig. 22 5E). The track angle histograms of empty-handed runs, composed of running 23 profiles on the unfamiliar floor, showed that angles larger than 120 degrees were 24 very frequent (lower panel in Fig. 5F). The profiles differed significantly between 25 the unfamiliar and familiar floors (P<0.01, Wilcoxon test). In the holding runs, most 26 of the track angles were smaller than 110 degrees until the workers stepped onto 27 the unfamiliar floor. Once a worker crossed onto the unfamiliar floor, even if they 28 ran back onto the familiar floor, angles larger than 120 degrees were very frequent 29 (lower panel in Fig. 5G). Track angles differed significantly before and after 30 stepping onto the unfamiliar floor (P<0.01, Wilcoxon test). 31 For all 43 workers that delivered all of five broods to the queen in both the 32 preliminary and final tests (Table 2B), there were significant differences between 33 the unfamiliar and familiar floor in both empty-handed and holding runs (P<0.01, 34 Wilcoxon test; Fig. 6). 35

9 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 [Discussion] 2 3 A Camponotus worker engaging in the brood-accumulation task alternates 4 between two different states, 'holding' and 'empty-handed'. In this study, I showed 5 that the responsiveness of a worker to the queen differs significantly between these 6 states, indicating that 'holding' represents a distinct physiological condition than 7 'empty-handed'. The motivation for holding appears to be finding a queen for the 8 brood being carried. This idea is supported by the observation that picking up a 9 brood is a cue to start searching for the queen. On the other hand, the 10 empty-handed state might also be associated with actual motivation. Immediately 11 after releasing the brood beside the queen (however, that is not a case at the first 12 brood), ‘empty-handed’ workers departed. Such quick conversion to the other state 13 might be explained by a motivation to search for broods not already adjacent to the 14 queen. Across insect species, internal representations of an animal’s own actions 15 have been shown to constitute building blocks for appropriate behaviors (cf. 16 Haberkern & Jayaramen, 2016). This idea supports the hypothesis that brood 17 accumulation behavior might be driven by alteration of internal representations 18 associated with the holding and empty-handed states. 19 For workers engaging in the brood-accumulation task, reaching the queen is 20 essential. To address this issue, I presented workers engaging in the task with an 21 unfamiliar floor by increasing the distance between the broods and the queen. 22 Regardless of whether they were holding or empty-handed, the workers strayed on 23 the unfamiliar floor. If workers recognize the distance from 'start' to 'goal' and 24 exploit it as a clue to achieve their tasks, then they should stray on the latter half of 25 their trip. Consequently, information about running distance would not be available 26 for brood-accumulation behavior. 27 Following chemical signposts is the most straightforward method for 28 reaching the destination. Pheromone trails are well-studied systems used to guide 29 foragers between the nest and food sources. Previous research on pharaoh’s ants 30 (Monomorium pharaonis) demonstrated that pheromone trails include polarity 31 information (Jackson et al., 2004). Inside the dark nest, multiple chemicals 32 including trails are attached to the tunnel surface because workers walk back and 33 forth within the narrow space. In addition, the queen moves and does not stay in the 34 same place. Under such circumstances, it would seem impossible for workers to 35 distinguish the polarity information included in the trails and exploit it to guide 36 their round-trip behavior. In this study, no workers quit their task because of the

10 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 unfamiliar floor (i.e., the presumed disappearance of trail pheromone). Thus, it 2 would be difficult to explain brood-accumulation behavior in terms of 3 route-following guided by pheromonal signals. 4 Marking is also well accepted as a type of chemical sign to operate in the 5 pinpoint. Home-range marking, for example, plays a crucial role in the vicinity of 6 the nest to define the property of the colony, and both indicate the nest entrance to 7 returning workers and provides defense against non-nestmates (Hölldobler and 8 Wilson, 1990). Such markings often consist of non-volatile compounds from the ant’s 9 cuticle surface, and therefore must be perceived by direct antennal contact (Lenoir 10 et al., 2009). This evidence supports the hypothesis that in order to achieve the 11 brood-accumulation task, workers must enhance orientation performance with the 12 help of some marking material that acts like a footprint. This idea is consistent with 13 this study’s finding that workers strayed, but never gave up on accomplishing the 14 task on the unfamiliar floor. 15 Recognition of and attraction to the queen by workers is generally mediated 16 by chemoreception (Hölldobler and Wilson, 1990). Therefore, the ability to select the 17 appropriate information from a complex mixture of chemical components is 18 important for the brood-accumulation task, as well as other social behaviors of 19 Camponotus workers. Attention is a physiological mechanism by which animals 20 select useful information by filtering the stimuli from the environment (Aston-Jones 21 et al., 1999). Human studies of attention have shown that this selection is controlled 22 either in a voluntary, top-down way or in a passive, bottom-up way (Motter, 1994; 23 Theeuwes, 2010). Recent studies in several insects have revealed attention-like 24 processes involved in selection of one among many visual and auditory stimuli (van 25 Swindern and Greenspan, 2003; Spaethe et al., 2006; Wiederman and O' Carroll, 26 2013; de Bivort and van Swindern, 2016), and careful comparisons of the 27 mechanistic similarities between insects and primates have been carried out (cf. 28 Nityananda, 2016). Very little work, however, has been published on olfactory 29 attention. In this study, I showed that brood-holding can lead a worker to pay 30 attention to the queen's label. It remains to be demonstrated whether this process is 31 elicited by attentional capture or an endogenous cueing procedure. Mature workers 32 that are isolated from the queen immediately after their emergence do not perform 33 the brood-accumulation behavior (Hara, 2003), indicating that 'social association' 34 might be necessary to induce some neuronal modulations associated with the 35 motivation for and/or the alteration of the holding and empty-handed states. To 36 address the mechanism of olfactory attention, it might be useful to exogenously

11 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 induce brood-accumulation behavior in such isolated workers. 2

12 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 [Acknowledgments] 2 3 I wish to express my deepest thanks to three undergraduate members in my 4 laboratory, Chie Sato (2005-2007), Sho Suzuki (2011-2013) and Ayane Tanaka 5 (2012-2014), for invaluable assistance with the collection of data. In fifteen years 6 from 2003 to 2017, a total of fifteen undergraduate students belonging to my 7 laboratory helped me to take care of the ant colonies. I also thank them. This study 8 was supported financially by Tokyo Gakugei University.

13 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 [References] 2 3 Aston-Jones GS, Desimone R, Driver J, Luck SJ, Posner MI (1999) Attention. In: 4 Zigmond MJ, Bloom FE, Landys, SC, Roberts JL, Squire L (eds.), Fundamental 5 Neuroscience. Academic Press, SanDiego, 1385-1409. 6 7 Collett M, Collett TS, Wehner R (1999) Calibration of vector navigation in desert 8 ants. Curr Biol. 9: 1031-1034. 9 10 de Bivort BL, van Swinderen B (2016) Evidence for selective attention in the insect 11 brain. Curr Opin Insect Sci. 15: 9-15. 12 13 Haberkern H, Jayaraman V (2016) Studying small brains to understand the 14 building blocks of cognition. Curr Opin Neurobiol. 37: 59-65. 15 16 Hara K (2002) A sensitive and reliable assay for queen discrimination ability in 17 laboratory-reared workers of the ant Camponotus japonicus. Zoolog Sci. 19: 18 1019-1025. 19 20 Hara K (2003) Queen discrimination ability of ant workers (Camponotus japonicus) 21 coincides with brain maturation. Brain Behav Evol. 62: 56-64 22 23 Hara K (2013) Social association brings out the altruism in an ant. In: Witzany G 24 (ed) Biocommunication of Animals. Springer, Capter9, 149-160 25 26 Holman L (2010) Queen pheromones: The chemical crown governing insect social 27 life. Commun Integr Biol. 3(6): 558-60. 28 29 Hölldobler B, Wilson EO (1990) The ants. The Belknap Press of Harvard University 30 Press, Cambridge, Mass 732 pp. 31 32 Ishii Y, Kubota K, Hara K (2005) Postembryonic development of the mushroom 33 bodies in the ant, Camponotus japonicus. Zoolog Sci. 22: 743-753. 34 35 Jackson DE, Holcombe M, Ratnieks FL (2004) Trail geometry gives polarity to ant 36 foraging networks. Nature 432: 907-909.

14 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 LenoirA, Depickère S, Devers S, Christidès JP, Detrain C (2009) Hydrocarbons in 2 the ant Lasius niger: From the cuticle to the nest and home range marking. J Chem 3 Ecol. 35: 913–921. 4 5 Merkle T, Wehner R (2008) Landmark cues can change the motivational state of 6 desert ant foragers. J Comp Physiol A 194(4): 395-403. 7 8 Motter BC (1994) Neural correlates of attentive selection for color or luminance in 9 extrastriate area V4. J Neurosci. 14: 2178-89. 10 11 Motais de Narbonne M, van Zweden JS, Bello JE, Wenseleers T, Millar JG, 12 d'Ettorre P (2016) Biological activity of the enantiomers of 3-methylhentriacontane, 13 a queen pheromone of the ant Lasius niger. J Exp Biol. 219: 1632-8. 14 15 Nityananda V (2016) Attention-like processes in insects. Proc Biol Sci. 283(1842): 16 20161986. 17 18 Spaethe J, Tautz J, Chittka L (2006) Do honeybees detect colour targets using serial 19 or parallel visual search? J Exp Biol. 209: 987-993. 20 21 Theeuwes J (2010) Top-down and bottom-up control of visual selection. Acta Psychol. 22 135: 77-99. 23 24 van Swinderen B and Greenspan RJ (2003) Salience modulates 20-30 Hz brain 25 activity in Drosophila. Nat Neurosci. 6(6): 579-86. 26 27 Wehner R (2003) Desert ant navigation: how miniature brains solve complex tasks. 28 J Comp Physiol A 189: 579–588. 29 30 Wiederman SD, O’Carroll DC (2013) Selective Attention in an Insect Visual Neuron. 31 Curr Biol. 23: 156-161. 32 33

15 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 [Figure legends] 2 3 Figure 1 4 5 Schematic view of the experimental setups for (A) Experiment 1 and (B) 6 Experiment 2. (A) A worker is allowed to carry a brood toward either the foster 7 queen (in the one of Q1 or Q2) or an unfamiliar queen (in the other). The red dots 8 indicate the datum-points for the array analysis. (B) A worker is allowed to carry 9 broods in the dish "B" toward the queen in the dich "Q" in the preliminary test. In 10 the middle of the final test, the Q-B distance is expanded by pulling the dish "Q" 11 with a thread to the position of Q'. Subsequently, the worker has to run across the 12 suddenly appeared (i.e., unfamiliar) floor to complete the task. 13 14 Figure 2 15 16 Behaviors in the central space in Experiment 1. Activities of a worker (A) during 17 holding run for the 3rd brood and (B) during subsequent empty-handed run. 18 g=arrival point; s=starting point. For (C) the holding runs and (D) the 19 empty-handed runs in one brood-accumulation task, tracks (upper panels) and the 20 array analyses (lower panels) of a worker. The colors indicate different chains of 21 running. FQ, the foster queen; UQ, an unfamiliar queen; B, the broods. 22 23 Figure 3 24 25 Results of array analyses from 77 workers in Experiment 1. (A) Box-plot 26 representation of the spatial activities per task. Mean±s.d. = 11.91±3.56 in the 27 holding runs, and 18.99±3.31 in the empty-handed runs. (B) The proportion of time 28 for the holding runs in the central space per task. Mean±s.d. = 0.20±0.15. *P<0.01. 29 30 Figure 4 31 32 (A) The population of time for the workers to stay in the room "FQ" to the duration 33 for the task. The data are presented as box plots every delivery. Because of shorter 34 stays in each, the data at the 3rd, 4th and 5th deliveries are combined in one. 35 Mean±s.d. = 0.33±0.21 at the 1st delivery, 0.10±0.09 at the 2nd and 0.05±0.05 at the 3, 36 4&5th. (B) Cumulative time per task to stay in the room "UQ". N, the number of

16 bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 workers stepping in the room "UQ"; #, the number of workers stepping in the room 2 "UQ" more than once at the 3rd, 4th or 5th deliveries. *P<0.01. 3 4 Figure 5 5 6 Behaviors of a worker in Experiment 2. (A) Tracks from beginning of the 7 preliminary test to the ‘release’ of the 3rd brood in the final test. The colors indicate 8 different chains of running. (B) A track of the holding run for the 3rd brood in the 9 final test. B and Q denote the dishes (outlined) for the broods and the queen, 10 respectively. A black short-dashed line indicates the boundary between the familiar 11 and unfamiliar floors. (C) A track of the empty-handed run just after expanding the 12 B-Q distance. (D) A track of the holding run for the 4th brood in the final test. (E) 13 Measurement of track angle (see text). Track-angle profiles in (F) the empty-handed 14 run and (G) the holding run. Upper panels illustrate the trajectory of (C) or (D), 15 respectively. The colors indicate the differences between the familiar and unfamiliar 16 floor in each empty-handed or holding run. *P<0.01. 17 18 Figure 6 19 20 For each run, the percentage of workers whose angle profile is significantly different 21 between the familiar and unfamiliar floor (P<0.01). N=43. 22 23 Table 1 24 25 The workers tested in Experiment 1. (A) Their responses to brood. (B) 26 Brood-accumulation behaviors of the workers identified as 'transportation'. 27 28 Table 2 29 30 The workers tested in Experiment 2. (A) Their responses to brood. (1), the number of 31 workers analyzed in the preliminary test; (2), only the workers identified as 32 'transportation' in the preliminary test are tested in the final test. (B) 33 Brood-accumulation behaviors of the workers engaging in 'transformation' in the final 34 test.

Fig1

A 2.7cm

central Q1 space Q2

B Q Q’ thread

5cm 5cm

Familiar floor Unfamiliar floor

Figure 1

Schematic view of the experimental setups for (A) Experiment 1 and (B) Experiment 2. (A) A worker is allowed to carry a brood toward either the foster queen (in the one of Q1 or Q2) or an unfamiliar queen (in the other). The red dots indicate the datum-points for the array analysis. (B) A worker is allowed to carry broods in the dish "B" toward the queen in the dich "Q" in the preliminary test. In the middle of the final test, the Q-B distance is expanded by pulling the dish "Q" with a thread to the position of Q'. Subsequently, the worker has to run across the suddenly appeared (i.e., unfamiliar) floor to complete the task.

Hara, TGU. bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Table 1 A Response

day after standstill ignore gathering transportation total emergence

1-3 18 5 1 0 24

4-7 0 7 13 5 25

8-11 0 2 11 35 48

12-14 0 0 1 51 52

total 18 14 26 91 149

Brood-accumulation Queen repetition No. of workers recognition correct 77 Yes erroneous 3 correct 11 No erroneous 0 total 91

Table 1

The workers tested in Experiment 1. (A) Their responses to brood. (B) Brood- accumulation behaviors of the workers identified as 'transportation'.

Fig2

A holding B empty-handed FQ FQ g s UQ UQ s g B B

C holding D empty-handed

Figure 2

Behaviors in the central space in Experiment 1. Activities of a worker (A) during holding run for the 3rd brood and (B) during subsequent empty-handed run. g=arrival point; s=starting point. For (C) the holding runs and (D) the empty- handed runs in one brood-accumulation task, tracks (upper panels) and the array analyses (lower panels) of a worker. The colors indicate different chains of running. FQ, the foster queen; UQ, an unfamiliar queen; B, the broods.

Hara, TGU. *P<0.01. *P<0.01. Mean space task. in per the central runs the holding time for and 18.99 runs, holding activitiesthe spatial representation of per Mean task. Results analyses Experiment1. Box-plot 77 (A)ofin from array workers Figure

bioRxiv preprint A

number of brocks

not certifiedbypeerreview)istheauthor/funder.Allrightsreserved.Noreuseallowedwithoutpermission. doi: holding https://doi.org/10.1101/383059

* empty-handed

± 3.31of proportionThe (B) empty-handed the runs.in

; this versionpostedAugust2,2018. B

proportion of duration ± s.d. = 11.91= s.d. The copyrightholderforthispreprint(whichwas holding ± s.d. = 0.20 = s.d.

± Hara, TGU. Hara, 3.56 the in ± 0.15. 0.15. Fig3

Fig4

A B * *

* proportion of duration

time time (s)

1st 2nd 3,4&5th

Figure 4

(A) The population of time for the workers to stay in the room "FQ" to the duration for the task. The data are presented as box plots every delivery. Because of shorter stays in each, the data at the 3rd, 4th and 5th deliveries are combined in one. Mean±s.d. = 0.33±0.21 at the 1st delivery, 0.10±0.09 at the 2nd and 0.05±0.05 at the 3, 4&5th. (B) Cumulative time per task to stay in the room "UQ". N, the number of workers stepping in the room "UQ"; #, the number of workers stepping in the room "UQ" more than once at the 3rd, 4th or 5th deliveries. *P<0.01.

Table 2 A Response

day after sample Behavioral standstill ignore gathering transportation emergence size (1) test

preliminary 0 2 4 25 8-11 31 (2) final 0 1 3 21

preliminary 0 0 1 36 12-14 37 (2) final 0 0 0 36 total total 68 57 in the final tests

Brood-accumulation

repetition mission No.of workers

complete 43 Yes incomplete 6

No ------8

total 57

Table 2

The workers tested in Experiment 2. (A) Their responses to brood. (1), the number of workers analyzed in the preliminary test; (2), only the workers identified as 'transportation' in the preliminary test are tested in the final test. (B) Brood- accumulation behaviors of the workers engaging in 'transformation' in the final test.

Fig5 A B B Q

C D E

F Empty-handed run G Holding run

Familiar Unfamiliar floor Familiar floor floor Unfamiliar floor

start

goal goal

start

* *

Figure 5

Behaviors of a worker in Experiment 2. (A) Tracks from beginning of the preliminary test to the ‘release’ of the 3rd brood in the final test. The colors indicate different chains of running. (B) A track of the holding run for the 3rd brood in the final test. B and Q denote the dishes (outlined) for the broods and the queen, respectively. A black short-dashed line indicates the boundary between the familiar and unfamiliar floors. (C) A track of the empty-handed run just after expanding the B-Q distance. (D) A track of the holding run for the 4th brood in the final test. (E) Measurement of track angle (see text). Track-angle profiles in (F) the empty-handed run and (G) the holding run. Upper panels illustrate the trajectory of (C) or (D), respectively. The colors indicate the differences between the familiar and unfamiliar floor in each empty-handed or holding run. *P<0.01. bioRxiv preprint doi: https://doi.org/10.1101/383059; this version posted August 2, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Fig6

Figure 6

For each run, the percentage of workers whose angle profile is significantly different between the familiar and unfamiliar floor (P<0.01). N=43.

Hara, TGU.