The black-legged tick Ixodes scapularis detects CO₂ without the Haller's organ

The survival and maturation of ticks require finding and blood-feeding on hosts. Both male and female ticks need at least one blood meal in each of their three development stages (larva, nymph, adult) to become sexually mature and reproduce (Apanaskevich et al., 2014). Especially in lay literature, the approximately 4% CO₂ of mammalian breath is often cited as a critical behavioral activator and attractant for ticks and an essential sensory cue for host seeking. However, few tick species have been tested for their CO₂ sensitivity, and quantitative laboratory studies are lacking.

CO₂ is a kairomone used by many hematophagous parasites to sense the proximity of a potential host (Chaisson and Hallem, 2012). For instance, CO₂ exposure elicits a range of contextual effects in various mosquito species that are thought to promote host-seeking behavior. These effects include activation of flight (Eiras and Jepson, 1991), increased attraction to body heat (Liu and Vosshall, 2019; McMeniman et al., 2014), and enhanced steering responses to vertically elongated dark visual objects (Vinauger et al., 2019). Whether CO₂ activates and promotes host seeking in ticks has not been well studied. However, there is evidence that exposure to CO₂ elicits behavioral responses related to host seeking.

The two dominant host-seeking strategies for ticks are active hunting and passive ambushing (Apanaskevich et al., 2014). Hunting ticks actively invade the habitat of hosts while ambushing ticks quest by climbing up the vegetation to wait for passing hosts, typically with extended forelegs. Many economically important tick species, including the black-legged tick Ixodes scapularis studied here, predominantly quest for hosts. It is well known among tick field ecologists that CO₂ traps efficiently capture tick species that hunt for hosts (Garcia, 1969, 1962; Guedes et al., 2012; Guglielmone et al., 1985; Norval et al., 1987; Oliveira et al., 2000). For example, the lone star tick Amblyomma americanum, known for its aggressive hunting behavior, is attracted by CO₂ traps from up to 21 m away (Wilson et al., 1972). However, species that predominantly quest to find hosts are only sporadically captured with CO₂ traps (Garcia, 1962; Guedes et al., 2012).

With about 850 species of ticks living in habitats from the cold Arctic to arid deserts, it is not surprising that anatomy, seasonal activity patterns, host preference and host-seeking strategies vary across species. For example, the winter tick, Dermacentor albipictus, is a so-called one-host tick that spends its life almost entirely on the same host (Howell, 1940). In contrast, I. scapularis is a three-host tick that feeds on a different host during each life stage and leaves the host to molt (Keirans et al., 1996; Soneshine and Roe, 2014). Larvae preferably feed on small mammals such as rodents, while nymphs and adults prefer larger hosts such as deer, dogs, horses and humans (Keirans et al., 1996). As a result of this three-host lifestyle, I. scapularis efficiently spreads blood-borne diseases such as Lyme borreliosis (Eisen and Eisen, 2018). How CO₂ contributes to host seeking or selection in this notorious disease vector is not understood.

Most studies examining CO₂ attractiveness for ticks have not aimed to explain the physiology of CO₂ detection. Instead, many just tested the effectiveness of CO₂ as bait and focused on quantifying changes in locomotion, i.e. counted how many ticks actively walked towards a trap (Garcia, 1969, 1965, 1962; Gray, 1985; Guedes et al., 2012; McMahon and Guerin, 2002; Miles, 1968; Nevill, 1964; Norval et al., 1987; Wilson et al., 1972). As tick species that predominantly quest for hosts are rarely captured with CO₂ traps, it is reasonable to ask whether the presence of CO₂ motivates questing ticks to walk and whether quantifying changes in locomotion is sufficient to describe CO₂-induced behavioral responses. Van Duijvendijk and colleagues (2017), for example, tested Ixodes ricinus nymphs, the European cousin of I. scapularis, for CO₂ attractiveness using a Y-maze choice olfactometer. Attractiveness was assessed by counting the number of nymphs that actively approached the odor arms. Possible non-walking-related behavioral reactions in response to CO₂ were not evaluated. Less than 25% of the animals tested walked during the selection test, and the few walking animals did not prefer CO₂ over air. The sparse number of walking ticks suggests that CO₂ might not induce walking in Ixodes ricinus nymphs. Questing larvae of D. albipictus, for instance, do not start to walk when exposed to CO₂ but remain in their questing position (Garcia, 1969).

The approach used here differs significantly from that of most previous studies in that it includes observations of behavioral reactions other than changes in locomotion. I demonstrate that CO₂ is a potent stimulant for adult male and female I. scapularis and that behavioral responses depend on the state of the animal. Walking animals tended to change their walking speed, while stationary animals remained stationary but started to wave their forelegs. This latter behavior was striking and resembled the described foreleg waving of questing ticks, assumed to support CO₂ detection. The forelegs of ticks carry a specialized multisensory structure, the Haller's organ (HO), which has been implicated in CO₂ detection (Carr et al., 2017; Steullet and Guerin, 1992). However, I provide strong evidence that the HO is not necessary for CO₂ detection and that other CO₂-sensitive structures not yet identified must be present on the tick body.

Pathogen-free unfed virgin adult male and female Ixodes scapularis Say 1821 ticks were purchased from Oklahoma State University Centralized Tick Rearing Facility (Stillwater, OK, USA) approximately 6 weeks post-molting. Upon arrival, the animals were housed for 4 weeks under controlled laboratory conditions prior to experiments. Ticks were housed in 60 ml aseptic screw cap plastic vials (VWR, 216-1822P) with a mesh-covered 25 mm round hole in the lid. Each vial contained a maximum of 10 ticks. Males and females were maintained in separate vials. The tick vials were kept in transparent, airtight food storage containers (5.4 l Clips & Close, Emsa, Emsdetten, Germany) at 99% relative humidity (saturated NaSO₄ solution). Food containers containing the tick vials were kept inside an incubator (KBW E5.1, Binder GmbH, Neckarsulm, Germany) with a photoperiod of 14 h light, 10 h darkness, and circadian temperature cycling (21°C, lights on 06:00–20:00 h; 18°C, lights off 20:00–06:00 h). The animals were used for experiments 10–18 weeks after molting, never received blood, and were always separated by sex. The data presented were collected on ticks from seven shipments over 18 months. There were no behavioral differences between ticks from different shipments, nor was there any seasonal variation in CO₂ responses of the lab-raised ticks used for this study.

All experiments were performed during the day between 10:00 h and 16:00 h at conditioned room temperature (21±0.5°C) and 40–60% relative humidity. For each experiment, five male or female I. scapularis were placed on the test area (30×30 cm) inside a custom-made airtight Plexiglas box (Fig. 1A, 40×40×30 cm) using a soft brush without prior immobilization. Ticks were not pre-selected by, for example, blowing on them and seeing if they started to walk – a procedure occasionally used for selecting ticks for choice experiments. The walking surface of the test area was lined with white vinyl foil (matte finish VViViD XPO, St-Laurent, QC, Canada) to ensure good contrast for video monitoring and a sufficient grip. The walking surface was cleaned with 70% ethanol before ticks were placed in the arena. A water moat prevented ticks from escaping from the test area.

Ticks were allowed to acclimate for 30 min before the experiments. To test tick responses to different CO₂ concentrations (Figs 1 and 2), I performed five experimental blocks on the same group of animals. Ticks were exposed to an air puff as a negative control and to 1%, 2%, 4% and 8% CO₂. The treatment order was not randomized in these experiments. Ticks were first exposed to the air puff, followed by ascending CO₂ treatments. CO₂ was increased globally in the arena and there was no detectable gradient (see ‘CO₂ measurements’, below). After acclimation, ticks were video monitored for 5 min in ambient air (2 frames s⁻¹) before the 5 min long treatment started. The animals were allowed to rest for 20 min between each experimental block, where the CO₂ concentration was reset to ambient air levels (350–600 ppm) by opening the arena door. In experiments where treatments were only an air puff and 4% CO₂ (Figs 3–5), stimuli were applied in randomized order.

Experiments were carried out in the same Plexiglas box as the multi-animal experiments, with identical equipment. The test area was replaced with a floating ball setup, modified after Loesche and Reiser (2021). To facilitate handling, single male or female ticks were briefly immobilized through cold treatment. Animals were then tethered to a metal rod for fixation. In short, immobilized ticks were picked up with a brush, placed into a movable cylindrical cavity (‘sarcophagus’) machined from solid brass, and positioned under a stereomicroscope (Stemi 2000, Zeiss). After aligning the tick in the sarcophagus, a thin metal rod was attached to the middle of the scutum with a small drop of UV-curable glue (UV liquid plastic, Bondic, Niagara Falls, NY, USA). The tethered tick was then mounted on a micromanipulator (MX10, Siskiyou) for precise positioning along the three translational axes on top of an air-supported ball (Last-A-Foam FR-7120, 20 mm diameter, General Plastics Manufacturing Company, Tacoma, WA, USA) that served as an omnidirectional treadmill. The air supporting the ball was carbon-filtered with 350–600 ppm CO₂. Care was taken that each tick was not cold immobilized for more than 5 min. Ticks were placed on the sphere in a horizontal position.

Experiments started after 30 min of acclimation at room temperature (21±0.5°C) and 40–60% relative humidity. Two experimental blocks were carried out on the respective individual animals. Ticks were either exposed to an air puff or to 4% CO₂ in randomized order and video-monitored for 4 min (5 frames s⁻¹, 2 min before and after treatment). Only the last minute in the ambient air and the first minute during the treatment were used for the analysis (Fig. 3).

To quickly change the CO₂ concentration inside the arena and with as little gradient as possible, 100% CO₂ was introduced into the Plexiglas box under high pressure (40 l min⁻¹). A diffusion pad (59-119, Genesee Scientific) attached to one of the side walls of the box ensured broad CO₂ distribution. The CO₂ inflow was timed with a solenoid valve (VZWD Series, Festo, Esslingen, Germany) controlled by a timing relay (LT4H Series, Panasonic). A flow meter (Q-Flow 140 Series, Vögtlin, Unna, Germany) was connected upstream of the solenoid valve to minimize disparity in CO₂ concentrations between experiments. A computer fan with speed control (12×12 cm, Wathai) next to the diffusion pad provided continuous air movement and CO₂ distribution. During the experiments, the fan was constantly running on low speed (∼0.2 m s⁻¹ airflow). This allowed sufficient airflow without creating wind and interfering with tick behavior.

CO₂ was measured with an infrared CO₂ Meter (model GM70+GMP251 Probe, Vaisala Inc.) at a sampling rate of 1 reading s⁻¹. To ensure that CO₂ was evenly distributed in the arena and that the measured CO₂ concentration corresponded to the concentration the ticks were experiencing, the CO₂ meter was placed at six different locations in the arena (on the four side walls, the ceiling and directly on the test area). For each CO₂ sampling location, three measurements were taken for each of the four CO₂ concentrations tested and measurements were compared (data not shown). Irrespective of the CO₂ concentration, measurements differed by ±50–250 ppm on average between sampling locations, which is within the measurement uncertainty of the CO₂ probe (±0.1%/1000 ppm at 5% CO₂; ±0.2%/2000 ppm at 8% CO₂) and was therefore ignored. During the experiments, CO₂ was measured at a mid-height at the rear side of the arena, with the probe placed slightly higher than the test area. CO₂ was measured in ppm, but is reported as percentage values throughout the paper. CO₂ delivery times were: 0.58 s for 10,000 ppm (1%), 1.25 s for 20,000 ppm (2%), 2.70 s for 40,000 ppm (4%), 5.75 s for 80,000 ppm (8%), each with a maximum deviation of ±500 ppm.

A puff of carbon-filtered compressed air was used as a negative control to distinguish possible CO₂-elicited behavioral changes from reactions to increased air movement during the onset of the treatments. The carbon-filtered air had the same pressure as the CO₂ (40 l min⁻¹) and was introduced the same way into the Plexiglas box. Air puff duration was 5.75 s, the same time period as for the 8% CO₂ treatment. Before the air puff treatment, hoses were flushed extensively with air to remove CO₂ residues. The CO₂ concentration of the carbon-filtered compressed air did not differ from that of ambient air (350–600 ppm).

It is questionable whether I. scapularis has a rudimentary vision. As a precaution, all experiments were performed in darkness under red LED lighting (640–670 nm, BestLEDStrip.com), a wavelength many arthropods cannot detect. Uniform lighting conditions were achieved by placing the LED lights around the test area. Ticks were video monitored with a monochrome camera (Blackfly S, FLIR) at 3.2 megapixels (2048×1536 pixels). The camera was positioned outside on top of the Plexiglas box and animals were monitored from above. Multi-animal experiments were recorded with a 6 mm/F1.4 fixed focal length lens (67-709, Edmund Optics) at 2 frames s⁻¹. Single-animal tick-on-a-stick experiments were recorded with a 94 mm macro lens (994100, Infinity) at 5 frames s⁻¹.

The following criteria were used to quantify behavioral changes in response to an air puff or CO₂ treatment (Figs 2, 4F and 5E).

(1) Stationary: animals that moved less than 5 mm within 2 min before the start of the treatments.

(2) Questing: stationary animals with raised forelegs for at least 1 min (example in Fig. 1B).

(3) Walking: animals that moved more than 5 mm within 2 min before the start of the treatments.

(4) Walking faster/slower: only applies to animals that showed a change in walking speed during exposure to an air puff or CO₂ compared with ambient air (control condition).

The mean walking speed (Δv_air) and standard deviation (s.d._air) over 60 s immediately before treatment were calculated. Faster walking comprises animals whose mean walking speed during the treatment (Δv_treatment) was greater than the mean walking speed plus standard deviation in ambient air (Δv_treatment>Δv_air+s.d._air). Slower walking comprises animals whose mean walking speed during the treatment was less than the mean walking speed minus standard deviation in ambient air (Δv_treatment<Δv_air–s.d._air). Walking animals that did not meet either criterion were further classified as walking.

In some experiments, the foreleg tarsi (Fig. 4; Fig. S4) and the mouth (Fig. S5F–I) were covered to test the contribution of the Haller's and the palpal organ to CO₂ detection. A paraffin-based casting wax (10815, Kerzenkiste, Bullenkuhlen, Germany) with a low melting point of ∼50°C was used. For a better visibility, the wax was dyed blue with liquid wax paint (10781-1-005, Kerzenkiste). To minimize the risk of tissue damage from heat, the wax was pre-melted in a wax warmer at 55°C (Nivlan wax warmer, Amazon). For application, the melted wax was picked up with a custom-made heating loop (diameter 0.6 mm, 0.19 mm silver wire) attached to a low-temperature soldering iron (WP80, Weller). This ensured that the wax did not harden between pickup and application. At least three layers were required to cover the long and wall-pored sensilla of the HO capsule completely. Cold immobilization and the brass sarcophagus used to glue ticks for the tick-on-a-stick experiments proved unsuitable for wax application because of condensation of atmospheric moisture on the cold surfaces and restricted access to the forelegs and mouthparts. It was thus necessary to immobilize ticks with 100% CO₂ using a diffusing pad (59-119, Genesee Scientific) positioned under a stereomicroscope (Stemi 2000, Zeiss). Care was taken that ticks were not exposed to the CO₂ for longer than 5 min. Ticks recovered within 1–2 min from the immobilization and were allowed to further recover in the incubator at 21°C and 99% relative humidity for 2 h before being tested. CO₂ immobilization had no detectable influence on tick behavior. Ticks immobilized with CO₂ for wax application on the foreleg patella (Fig. S5F–I) showed similar behavioral responses to 4% CO₂ to those of non-immobilized ticks (Fig. 1).

In some experiments, the foreleg tarsi and HO were amputated. Ticks were immobilized on CO₂ for amputation as described above, and the tarsi were cut off with spring scissors (Vannas Spring Scissors, FST). After amputation, the animals were allowed to recover in the incubator at 21°C and 99% relative humidity for 2 h before being tested.

To test whether the wax used to cover the foreleg HO and mouthparts is impermeable to CO₂, the CO₂ probe was coated with wax. One of the 60 ml tick housing plastic vials (216-1822P, VWR) with a 25 mm mesh-covered hole in the lid was taken and the lid was coated with a thin layer of the low melting point casting wax used to cover tick body parts by briefly dipping it into the melted wax in the wax warmer. A second hole was drilled into the bottom of the plastic vial where the CO₂ probe was inserted. A clipped balloon (pink wrap in Fig. S4D) and several layers of Parafilm were used to seal the hole where the CO₂ probe was inserted and the lid. The covered probe was then placed inside the Plexiglas behavior arena, and CO₂ was added to the box for 2.7 and 6 s (3 trials each), corresponding to a CO₂ concentration of 4% and 10%. CO₂ concentration was reset to ambient air levels between trials by opening the arena door. The same experiment was carried out with a non-wax-covered lid to rule out that the CO₂ measurements were affected by the attached plastic vial. A new lid was screwed onto the vial without affecting the balloon/Parafilm seal.

Reconstruction of the animals' x–y body and foreleg positions was achieved with the Tracker Video Analysis and Modeling Tool (Open-Source Physics, version 6.1.1). In multi-animal experiments, the center of the body (center of mass) was tracked; in single-animal experiments, the tip of the left and right forelegs was tracked. The tracked x–y coordinates were further processed and analyzed with custom-written MATLAB scripts (MathWorks, version R2023a).

Walking speed (mm s⁻¹) was calculated from x–y coordinates. Values were normalized to compare changes in walking speed across animals. For this, the data for the 10 min long experiments were binned (bin width=30 frames/15 s) by calculating the mean velocity for each bin and each animal. Data from each animal were normalized to the respective mean of the experiment's first 10 bins/150 s.

The coefficient of variation (CV) was calculated from the normalized binned data using CV=σ/μ, where μ is the mean of each bin and σ is the respective standard deviation. CV≠1 indicates the data deviated from the bin's mean value.

To quantify the number of ticks that started to raise their forelegs or body, manual video analysis was necessary. The analysis was performed blind, by encrypting the file names and selecting them randomly (ImageJ, Blind Analysis Tool). Two minutes pre- and post-treatment of the 10 min long recordings were analyzed.

Foreleg angular positions were determined by placing a virtual polar coordinate system through the basal part of the animals' mouthparts (base capitula, see Fig. 3A). The average body length was used to calibrate the x–y coordinate system in the Tracker software (tip of the mouth to alloscutum; females 4 mm, males 3 mm). The four-quadrant inverse tangent in degrees was calculated from tracked x–y coordinates: 0 deg corresponds to forelegs exactly parallel to the midline; negative angles are left and positive angles are right. For x–y foreleg positions (Fig. 3D,F), distances are given in mm: 0 mm corresponds to the axis intersection point at the basal part of the animal's mouthparts.

Position–time diagrams (Fig. 3B; Fig. S3) were used to detect the local maxima of angular positions of the left and right foreleg and the second pair of leg tips as a function of time. Peaks were considered separated if the minimum distance between adjacent peaks was at least 0.4 s or 2 frames. The movement frequency (Fig. 3C) was calculated from the total number of local maxima over 60 s before and after the onset of the treatment.

To compare initial walking speeds, the mean walking speed over 5 min in ambient air was calculated for all animals that walked. In contrast, to compare walking speeds during the 4% CO₂ treatment, only walking and faster walking animals were included and the mean walking speed over 5 min was calculated.

The probability density estimate of the angular distribution of the forelegs (Fig. 3E) was calculated and plotted with the Raincloud Plot extension for MATLAB (Allen et al., 2021). Total density of the foreleg tip x–y positions (Fig. 3F) was determined by reducing the x–y coordinate system to one axis (x and y, respectively). The total length of the axis (8 mm, −4 mm to +4 mm) was divided into 100×1 pixels (0.08 mm pixel⁻¹). For each pixel, the sum of the x–y positions of the left and right foreleg tip was calculated (the total density). Data were not normalized and are given in mm distance from the axes intersection (the basal part of the animal's mouthparts). The data were smoothed with a 5-pixel moving average filter.

Statistical analysis was performed with MATLAB (MathWorks, version R2023a). The Shapiro–Wilk test was used to test data for normal distribution. Data are reported as means±s.d. All data tested for statistical differences were normally distributed. Results from t-tests are reported as t_{degrees of freedom}=t-value, P-value either in the text or in the figure legends. Final figures were prepared with Adobe Illustrator (version 24.2.3).

To determine behavioral responses and sensitivity to CO₂, unfed adult I. scapularis were exposed to different CO₂ concentrations (1%, 2%, 4% and 8%). An air puff with the same duration as the 8% CO₂ treatment was used as a negative control to distinguish behavioral responses to CO₂ from changes in airflow. The treatment order was not randomized in these experiments: air puff controls were always first, followed by ascending CO₂ treatments. I tested 50 ticks, 25 males and females, in groups of five. In each experiment, five ticks were placed inside a Plexiglas arena (Fig. 1A) and video monitored for 10 min (5 min in ambient air and during treatments).

The animals were allowed to behave freely, and accordingly, different behavioral states in control conditions (ambient air) were observed: some animals walked, while others were stationary and did not move. I took an unbiased approach and looked for changes in behavior for walking and non-walking animals alike. The two most prominent behavioral reactions to CO₂ were changes in walking speed and extensive foreleg waving, often associated with changes in posture. Foreleg waving was easily identifiable in non-walking animals, whose forelegs typically pointed forward and rested on the ground in ambient air (Fig. 1B, top). In the presence of CO₂, many of these animals began to raise and wave their forelegs (Fig. 1B, middle; Movie 1). In some cases, animals would hold their entire body erect while waving with their forelegs (Fig. 1B, bottom). During walking, the forelegs were often pointed forward, constantly moving, and touching the ground at irregular intervals. The animals were video-monitored from above. Because of the extensive foreleg movement during walking, it was not possible to resolve alterations in foreleg position for walking animals in these experiments. It is likely, however, that CO₂ elicited increased foreleg waving in walking animals, too.

To quantify how many animals responded to the treatments, changes in walking speed and foreleg movements were scored with either 0 (no change) or 1 (behavioral change) over 2 min in ambient air and CO₂, respectively. Overall, the air puff did not strongly affect I. scapularis’ behavior. Only 16% of the animals responded (Fig. 1C). However, I. scapularis showed high sensitivity to all tested CO₂ concentrations: 80% of the 50 animals showed behavioral responses to 1% and 2% CO₂. Most behavioral changes were observed at 4% CO₂ (90%) and slightly less at 8% (82%). The data thus show that I. scapularis has no clear concentration preference in the tested range and is not more strongly tuned to breath-like CO₂ concentrations (∼4%) than to the other concentrations tested.

To test for sex-specific responses, I analyzed males and females separately (Fig. S1). No substantial differences were observed. Males and females responded similarly to the tested CO₂ concentrations (Fig. S1), with males showing slightly more reactions at 1% (88% versus 72% females) and females at 4% CO₂ (96% versus 84% males). Thus, male and female data were pooled in all further experiments.

Ixodes scapularis’ behavior was quite variable in the control condition (ambient air): animals displayed various behaviors that changed with time. For example, animals often did not walk continuously but paused at irregular intervals before resuming locomotion. Consequently, a simple comparison of walking speeds before and during treatment showed no difference, because of the high variability (Fig. S2A). However, I found sudden decreases in walking speed at the onset of the stimulus for all applied treatments, including the air puff (dark band in Fig. S2A). Manual video analysis revealed that many ticks stopped briefly and began to wave their forelegs. As these initial speed decreases were short and lasted roughly 10 s, they are referred to here as startle responses. Because 30% of the ticks showed startle responses during the control air puff (Fig. S2B), I hypothesize that these were partly mechanosensory and elicited by changes in airflow.

Despite the high variability, I aimed to quantify behavioral changes in walking speed in more detail. To detect possible changes, each animal's data were normalized to its average walking speed during the first 150 s of the experiment (10 bins). Thus, an increase in walking speed resulted in numbers greater than 1 and a decrease resulted in numbers smaller than 1. Fig. 1D shows the change in walking speed as heatmaps for the 50 animals tested throughout the 10 min experiments. The number of animals that increased their walking speed rose with higher CO₂ concentrations, as evident from a shift towards lighter colors in the heatmaps. However, some animals also decreased their walking speed or stopped entirely in response to CO₂ (darker colors in heatmaps). Note that reductions in walking speed are under-represented in the heatmap color space as they can only range between 1 (initial walking speed) and 0 (animal is stationary). In contrast, increases in walking speed ranged between 1 and 5 in the experiments.

To eliminate this imbalance, the CV was calculated from the normalized data (Fig. 1E). Any deviation from the initial walking speed will increase the CV, independently of whether the animal slowed down or walked faster. The control air puff caused no substantial effect, and CV progression over time was almost linear. In contrast, CVs for the tested CO₂ concentrations increased rapidly at CO₂ onset, with most of them reaching their maximum about 1 min after the start of the treatment.

To conclude, the data show that I. scapularis respond robustly to changes in ambient CO₂ levels, independent of the concentration. Adding either air or CO₂ to the arena triggered short-term startle responses, presumably because of a change in airflow. After the startle, CO₂ exposure generally altered tick behavior for the remainder of the experiment. By contrast, during air puffs, the ticks typically resumed the behavior they had shown before the puff.

A significant proportion of the animals did not walk and were stationary in the control condition. It is well known that an animal's behavioral state can influence the assessment of a sensory stimulus and associated behavioral responses. For example, fruit flies (Drosophila melanogaster) are attracted to CO₂ while flying but repelled by it when walking (Wasserman et al., 2013). A similar state dependency might exist in ticks and the way ticks respond to CO₂ might depend on whether the animals are walking or not.

To test whether the animal's physiological state affected their response to CO₂, data from walking and stationary ticks were separated based on the behavior displayed in ambient air (Fig. 2). Changes in walking speed and foreleg movement were analyzed 2 min before and after the start of treatment. Percentage values are reported rather than animal numbers because I noted a steady decrease in walking animals during the 3 h experiments while the number of stationary animals increased (Fig. 2A, middle column: 80% walking at the beginning, 42% at the end). As treatment order was not randomized in these experiments, the shift from walking to stationary is most likely caused by incipient dehydration over the experiment time. However, CO₂ responsiveness was not affected by this shift. It is conceivable that the physiological state of the animals changed during the experiment, which could influence the numbers presented, such that slightly more stationary animals will start to walk at higher CO₂ concentrations.

Exposure to CO₂ clearly induced behavioral changes in stationary and walking animals. Surprisingly, CO₂ did not induce robust walking in stationary animals. Most stationary animals remained stationary when exposed to CO₂ but started to wave their forelegs (Fig. 2A, right column, ‘questing’). Host-seeking I. scapularis often quest by climbing a patch of grass, stretching out, and waving their forelegs while waiting for passing hosts (Keirans et al., 1996). Because of the similarity to the natural behavior, increased foreleg waving and posture changes in stationary animals are referred to here as questing. Each CO₂ concentration tested elicited questing in stationary animals, with a peak at 4% CO₂ (93%) and a minimum at 1% CO₂ (75%). Ticks often remained in the questing position until CO₂ levels were reset to ambient air levels at the end of an experimental block. This long-lasting foreleg waving was exclusively observed during CO₂ exposure and differed from the brief startle responses. None of the animals started to quest in response to the air puff (Fig. 2Bi). Only a few stationary animals began to walk (Fig. 2A, right column), with the highest percentage (25%) at 1% CO₂ (Fig. 2Bi).

The response of walking animals was markedly different: most walking animals continued to walk in the presence of CO₂ (Fig. 2A, left column, sum of walking, walking faster and walking slower), and walking speeds differed substantially between the air puff and CO₂ exposure. During the air puff, most ticks showed no change in walking speed, but in CO₂, most ticks either slowed down or sped up. The highest percentage of slower walking animals was observed at 1% CO₂ (16%, Fig. 2Bii). Faster walking was most pronounced at 4% CO₂ (36%), suggesting that higher CO₂ concentrations promote faster walking. Independent of the CO₂ concentration, approximately one-third of the walking animals became stationary, and most of these stationary animals started to quest (Fig. 2A, left column). This starkly contrasted the responses to air puffs, where only a tiny percentage of the animals stopped walking, and those did not quest (Fig. 2Bii). Given that walking speed changed when CO₂ was applied, I suspected that walking direction and the number of directional changes were also affected. However, I found no consistent pattern in the number of directional changes, turning direction and magnitude, angular acceleration and walking direction, either in the control conditions or in CO₂ (data not shown).

In summary, CO₂ behavioral reactions were state dependent in that stationary animals started to quest, while walking animals continued to walk but at different speeds. The number of animals that increased their walking speed rose with the CO₂ concentration and reached its maximum at 4% CO₂. Questing in stationary animals was pronounced at all tested CO₂ concentrations but strongest at 4%.

The data indicate that CO₂ is sufficient to trigger foreleg waving in I. scapularis (Fig. 1B,C). To characterize CO₂-induced changes in foreleg waving in more detail, I performed single-animal experiments by exposing 8 I. scapularis individually to 4% CO₂. Ticks were tethered to a metal rod and positioned on an air-supported ball (Fig. 3A; Movie 2). The ball acted as an omnidirectional treadmill, allowing the animals to walk without changing position. As before, an air puff was used as a negative control, and animals were video-monitored from above.

The angular position of the forelegs was analyzed over 120 s (60 s each in ambient air and CO₂) by placing a virtual coordinate system through the basal part of the animal's mouthparts (Fig. 3A, inset). Fig. 3Bi,ii shows an example of this analysis for one adult walking female during an air puff and exposure to 4% CO₂. The plots on the left show the angles of the forelegs over time, while the images on the right show the position of the forelegs in relation to the animal. Fig. S3 shows the forelimb movement of all 8 animals tested. During CO₂ exposure, two noticeable changes occurred: the waving frequency increased, and the animal spread its forelegs further. This is evident by larger oscillations and angles in the red traces in Fig. 2Bii.

None of the tested animals were stationary in these experiments. All animals walked continuously or intermittently in control ambient air and during treatments. The air-supported ball appeared to have motivated the animals to walk, which was beneficial to the study because it revealed that CO₂ induces foreleg waving in walking ticks, too. Animals did not consistently use their forelegs for locomotion but instead appeared to use them for active sensing. This is based on the observation that in some instances the gait switched from eight-legged walking in ambient air to six-legged walking with extended and rapidly moving forelegs in CO₂. Active sensing is a common phenomenon in arthropods, akin to sniffing in mammals. In insects, movement of the usually bilaterally symmetric sensory appendages during active sensing is often coordinated and coupled to leg movements (Dürr et al., 2001; Horseman et al., 1997). However, the movements of the left and right forelegs were not strongly synchronized in these experiments. There were brief synchronized phases in which the left and right legs moved approximately parallel towards each other (left–right waving) or in opposite directions (inward–outward waving) and uncoordinated phases in which the left and right legs moved independently (Fig. S3). This suggested that, at least at times, the coupling between the left and right foreleg was weak or absent and that legs can be independently controlled for active sensing.

To quantify the effect of CO₂ on foreleg waving, the movement frequency over 60 s before and after the start of treatment was measured. The movement frequency of the second pair of legs was also analyzed to determine whether potential gait switches from six- to eight-legged walking influenced foreleg waving. Fig. 3C (left) shows that foreleg movement frequency during the air puff did not differ significantly from that in ambient air (mean±s.d. ambient air: 0.71±0.16 Hz, air puff: 0.76±0.16 Hz; P=0.50). By contrast, during 4% CO₂ exposure, foreleg movement frequency increased in 7 of the 8 animals tested, and the increase was significant compared with ambient air (ambient air: 0.67±0.14 Hz, 4% CO₂: 0.92±0.14 Hz; P=0.01). By contrast, the movement frequency of the second pair of legs in ambient air did not change significantly from that during the air puff (ambient air: 0.56±0.14 Hz, air puff: 0.58±0.12 Hz; P=0.49) or the 4% CO₂ treatment (ambient air: 0.57±0.12 Hz, 4% CO₂: 0.60±0.11 Hz; P=0.51). The significant increase in foreleg movement frequency during the 4% CO₂ treatment was independent of potential gait switches or changes in walking speed.

The spread of the forelegs was also consistent. The heatmaps in Fig. 3D depict the position of the forelegs for all 8 animals tested. In ambient air and during the air puff, the animals moved their forelegs within a range of approximately ±90 deg from the midline, with the highest density at ±40 to ±50 deg (Fig. 3E). In the presence of 4% CO₂, the range of motion increased to approximately ±100 deg from the midline, with the highest density at about ±70 deg. On average, foreleg motion shifted 0.4–0.8 mm outwards from the midline and 1.2 mm posteriorly (Fig. 3F).

The mechanisms of tick chemoreception are unknown, but they are mainly ascribed to the foreleg HO (Carr et al., 2017; Gebremedhin et al., 2023). The HO is unique to ticks and mites and presumed to function like the insect antennae. The structure and organization of the HO varies across tick species (Josek et al., 2018). Generally, the HO is located on the dorsal surface of the tarsus of each foreleg. It includes four functionally and structurally different parts: a capsule, an anterior pit, the distal knoll and a post-capsular area. Each part contains a definite number of structurally different sensilla. The morphology of the HO in adult I. scapularis is insufficiently understood. The only available examples come from transmission electron microscopy of the HO of Ixodes persulcatus (Leonovich, 1977) and light microscopy studies of I. ricinus (Leonovich and Belozerov, 1992). In these species, the capsule contains seven wall-pored sensilla, which are presumed olfactory based on their morphology. The anterior pit contains six sensilla. One is wall pored and innervated by six bipolar sensory neurons. It is unclear how many sensilla the posterior capsule of the genus Ixodes contains, as it is covered by cuticle except for a slit and is difficult to access. However, the anterior pit of Amblyomma variegatum bears seven sensilla. Three of these sensilla are wall pored and CO₂ sensitive (Steullet and Guerin, 1992).

Although not previously described in detail and not well documented in the literature, I am not the first to notice that CO₂ elicits foreleg waving in certain tick species (McMahon and Guerin, 2002). It is, therefore, widely believed that the HO is the main sense organ for detecting CO₂. To my knowledge, however, no study has carefully tested this hypothesis. I thus tested ticks with disabled HO for CO₂ responses by covering the foreleg tarsi with low melting point wax.

Fig. 4A shows an adult female with wax-covered HO. First, I confirmed that the wax was impermeable to CO₂. For this, a CO₂ probe was covered with a thin layer of wax (Fig. S4D), about the same thickness as the wax applied on the forelegs. The covered probe was placed inside the behavior arena, and CO₂ was added for 2.7 and 6 s. If the wax is impermeable to CO₂, measurements should stay the same once CO₂ is present. Fig. 4B shows that this was the case. Measured CO₂ concentration in the arena increased only by 0.015% (150 ppm) on average for 2.7 and 6 s of CO₂ application, respectively. This amount is within the measurement uncertainty of the probe. For comparison, 2.7 and 4 s CO₂ application resulted in a measured concentration of 4% and 10%, respectively, when the probe was not covered with wax.

I then tested I. scapularis' responses in the multi-animal walking arena (Fig. 1A) to CO₂ application with the HO covered with wax. One hundred ticks in groups of 5 were tested. Responses were only tested to 4% CO₂, as I. scapularis did not show a strong concentration preference (Fig. 1C). I first noted that ticks with covered HO rarely startled at the beginning of the treatment, evident by the missing dark band at the treatment onset (Fig. S4A). Only 8% of the 100 ticks with covered HO decreased their walking speed at the onset of the air puff, and 6% for the 4% CO₂ treatment (Fig. S4C). For comparison, the air puff triggered startle responses in 30% of the animals with intact HO and in 24% for the 4% CO₂ treatment. The decrease in startle responses shows that the wax coating was successful, and that the HO appears to detect changes in airflow that prompt startle responses.

As for ticks with intact HO, the air puff did not strongly alter I. scapularis’ behavior throughout the trial. Only 18% of the animals with covered forelegs showed changes in behavior (Fig. 4C; Movie 1). Surprisingly, however, 60% of the animals responded to 4% CO₂, although the HO was covered. Likewise, a comparison of normalized walking speeds in ambient air and during treatment revealed that walking speeds increased in response to 4% CO₂. The increase in walking speed is visible in the shift to lighter colors in the heatmaps (Fig. 4D) and the increase in CV during the 4% CO₂ treatment (Fig. 4E). Additionally, I noted that ticks with covered HO walked significantly slower than those with intact HO, even in ambient air (intact: 1.0±0.7 mm s⁻¹, N=22; covered: 0.5±0.4 mm s⁻¹, N=49; t-test: t₆₉=4.712, P<0.001). However, the wax application did not interfere with I. scapularis’ ability to walk – walking was just slower. Although walking speed in ambient air was initially decreased in ticks with covered HO, this difference disappeared during CO₂ exposure. Most ticks that changed their walking speed in the presence of 4% CO₂ started to walk faster (Fig. 4F, left panel). They reached speeds similar to those of ticks with intact HO (intact: 1.6±1 mm s⁻¹, N=13; covered: 1.5±0.7 mm s⁻¹, N=40; t-test: t₅₁=0.247, P=0.806).

Covering the HO had further effects. First, fewer animals started to quest when the HO was covered. Only two walking ticks became stationary, with one starting to quest during 4% CO₂ exposure (Fig. 4F, left panel). Questing was still present in stationary animals, albeit less prevalent than in animals with intact HO (Fig. 4F, right panel). Concurrently, the percentage of stationary animals that remained stationary without questing (Fig. 4F, right panel) was increased. Second, questing duration was reduced. Untreated animals often quested until the CO₂ concentration was restored to ambient air levels at the end of an experimental block (5 min and longer). In contrast, animals with covered HO only quested for 157±67 s on average (N=23). Third, the latency between the start of the 4% CO₂ treatment and the onset of the behavioral reaction was significantly increased. Untreated ticks with intact HO showed behavioral changes after 3.8±2.5 s (N=45). In comparison, ticks with covered HO showed behavioral changes after 17.1±7.2 s (N=60, t-test: t₁₀₃=−11.899, P<0.001 compared with 4% CO₂ intact HO).

I excluded that the additional weight of the wax caused the decrease in questing by applying a comparable amount of wax to the patella proximal on the forelegs (Fig. S5A–E). The additional weight on the patella did not strongly impact CO₂ reactions. A similar number of ticks with wax on the patella and untreated ticks with intact HO responded to 4% CO₂ (84% versus 90%). In addition, all stationary animals that remained stationary started to quest and the onset of behavioral changes was not statistically different from that for untreated animals (wax on patella versus intact HO with 4% CO₂: 4.1±3.2 s, N=42, t-test: t₈₅=−0.436, P=0.664).

I conclude that the HO is not necessary for CO₂ detection. However, CO₂ induces foreleg waving, supporting that CO₂ acts as an activator and primes the animals to quest for additional sensory cues. Fewer ticks started questing, and questing time was reduced when the HO was covered. This suggests that the animals lose interest in CO₂ more quickly when no active sensory input from the HO exists.

The data presented here contradict the assumption that I. scapularis HO is the primary CO₂ detector. Therefore, as an additional control experiment, the foreleg tarsi and the HO were amputated for 50 ticks, and responses were tested to 4% CO₂ and the air puff (Fig. 5A; Movie 1). The amputation affected I. scapularis behavior, especially in the smaller males. Seven of the 25 amputated males died within 24 h after amputation, while all females survived. As before, ticks with amputated forelegs walked slower in ambient air (0.4±0.5 mm s⁻¹, N=22) than walking animals with intact HO. Regardless of the animals' potentially impaired condition, 58% of the animals responded to 4% CO₂ either by changing their walking speed or by starting to quest (Fig. 5B–D). However, walking animals that continued walking in CO₂ did not reach the same average speed as ticks with covered HO or untreated animals (0.9±0.3 mm s⁻¹ in comparison to ∼1.5 mm s⁻¹, N=14). Overall, more animals decreased their walking speed in response to CO₂ (Fig. 5E, left panel), most likely as a result of the amputation. Similar to ticks with covered HO, fewer animals with amputated forelegs started to quest in response to 4% CO₂ (Fig. 5E), and questing time was reduced (169±46 s, N=13). In addition, the latency between the start of the 4% CO₂ treatment and the onset of a behavioral reaction was significantly increased compared with the response of ticks with intact HO to 4% CO₂ (17.9±7.2 s, N=29, t-test: t₇₂=−12.065, P<0.001).

In search of the structure mediating CO₂ detection, I covered the palpal organ with wax in a subset of experiments in addition to the HO. The palpal organ is a sensory structure on the tick mouthparts, presumed to be involved in chemosensation. Several multi-innervated sensilla have been identified in A. americanum that either have an elaborate pore canal system or have a single slit opening at the tip (Foelix and Axtell, 1971). However, covering the palpal organ and its sensilla did not prevent I. scapularis from reacting to CO₂: 65% of the 20 animals tested changed their walking speed or started to wave their forelegs in response to 4% CO₂ (Fig. S5H). This suggests that there are other sensory structures, besides the palpal and HO, mediate CO₂ responses in I. scapularis that still need to be identified.

Taken together, ticks with amputated and covered HO reacted similarly robustly to CO₂, providing further evidence that the HO is not necessary for CO₂ detection.

In summary, adult male and female I. scapularis responded robustly to CO₂ in the tested concentration range of 1–8%. Notably, CO₂ responses were maintained even when the HO was disabled, either through covering it with wax or by amputating the foreleg tarsi.

Ticks did not show a clear concentration preference. While the highest number of behavioral responses was observed during exposure to a breath-like concentration of 4% CO₂, all tested CO₂ concentrations elicited robust behavioral responses with only marginal differences (Fig. 1C).

The lack of a strong CO₂ concentration preference may seem counterintuitive. However, a concentration preference is not required to locate a CO₂ source. Odors are not distributed in a laminar manner in the air but rather as turbulent plumes of fine odor filaments with defined edges (Rigolli et al., 2022). The spatiotemporal pattern of these filaments depends on the distance from the odor source. Concentrations are typically highest at the origin and decrease with distance, while the odor plume becomes increasingly patchy and intermittent. While turbulent wind currents in the environment carry odorants from a source to an animal's olfactory organ (e.g. antenna, nose, forelegs), small-scale currents near the organ surface and molecular diffusion carry odorants to the olfactory receptors (Ache et al., 2016). Many animals actively interact with these small-scale currents, for example by sniffing or increased movement of sensory structures such as antennae or forelegs, also known as active sensing (Crimaldi et al., 2022).

Given the turbulent structure of odor plumes, detecting and staying within the edges of an odor plume is a more reliable detection strategy than using the concentration gradient itself. To detect edges, noticing CO₂ at low concentrations is essential. Mosquitoes, for example, track their hosts over long distances – up to 50 m in some cases – by following host-emitted cues such as CO₂ (Dekker et al., 2005; Healy and Copland, 1995). Many mosquito species can detect even small increases in the ambient CO₂ concentration. In Aedes aegypti mosquitoes, CO₂-sensitive maxillary palp sensilla neurons increase their action potential firing rate as soon as the CO₂ concentration is slightly above ambient levels (Grant and O'Connell, 2007). Whether I. scapularis have a similarly low CO₂ sensitivity is unknown. However, a study using the tropical bont tick A. variegatum has demonstrated that sensilla in the HO show threshold CO₂ responses between 0.002% and 0.1% (Steullet and Guerin, 1992). This suggests that, at least in principle, A. variegatum can detect the CO₂ plume of relatively distant hosts. Whether this is also the case for I. scapularis remains to be tested. However, in the present experiments, 80% of the animals tested reacted to 1% CO₂ exposure (Fig. 1C), suggesting that detection of CO₂ at concentrations below 1% is likely. Although the CO₂ concentrations tested here probably only occur at close host distances, the results indicate that CO₂ can serve as a long-distance host cue and that hungry adult I. scapularis become aware of approaching hosts early on. Detecting hosts from a greater distance is beneficial as it gives the animals more time to gather additional information and decide whether attaching is rewarding.

Host seeking is not a single behavior but rather a sequence of behaviors. A typical host-seeking sequence in endoparasites and ectoparasites alike is (1) activation of the animal, (2) orientation towards the potential host and, finally, (3) host selection and invasion. The expression of each of these phases varies between different animal species. Thus, it is sensible to assume that orientation behaviors differ between ticks that primarily quest for hosts and ticks that actively hunt. The data presented here show that CO₂ activates I. scapularis and initiates behaviors that resemble host seeking. Specifically, I identified two behavioral responses that depended on the state of the animal and can readily be assessed in laboratory conditions: walking ticks primarily increased their speed, mirroring a faster approach to a host during active hunting; in contrast, stationary ticks raised their forelegs and began to wave (Fig. 2). Foreleg waving strongly resembles I. scapularis' questing behavior in their natural habitat. It also mirrors the antennae and forelimb movements of other arthropods during active sensing. Active sensing is widespread in animals and is used, among other things, to recognize additional sensory cues and to sample the environment.

CO₂, by itself, is an ambiguous sensory stimulus. It is a by-product of plant respiration, decomposition of plant and animal matter, and the burning of oil, coal and gas. Thus, a CO₂ plume encountered in the environment can have vastly different behavioral relevance. Approaching a non-host CO₂ source can endanger the animal or, at best, waste already scarce energy resources. Information about the context of the CO₂ presence is thus essential to make optimal decisions. Breathing animals provide visual targets, create thermal gradients and release hundreds of compounds from the skin and in the breath, all contributing to a species-specific multimodal sensory stimulus. It is thus likely that ticks use strategies similar to parasitic nematodes and mosquitoes (Hallem et al., 2011; McMeniman et al., 2014; Vinauger et al., 2019), where CO₂ enhances the attraction to other sensory stimuli, such as host-specific odors and body temperatures resembling vertebrate body heat.

The forelegs of ticks carry the HO, a multisensory organ implicated in olfaction (Carr et al., 2017; Josek et al., 2021) and thermotaxis (Carr and Salgado, 2019; Mitchell et al., 2017) in certain tick species. Although direct evidence in I. scapularis is lacking, it is conceivable that there are also mechanosensory sensilla on the foreleg tarsi. I suggest that the CO₂-evoked waving of the forelegs in stationary and walking animals (Figs 2 and 3) is a form of active sensing and allows the detection of additional (multimodal) stimuli. In fact, I observed that walking animals that moved their forelegs during walking significantly increased the speed and range of waving movements in CO₂. This is consistent with the notion that the increased waving of the forelegs is based on the need to increase the capture rate of odorants, for example. I also observed that ticks occasionally erected their entire body while waving their forelegs. To effectively detect potential air-born or thermal stimuli, it should not make a difference for the animals whether they are erect or not. Most odor-tracking arthropods do not erect their bodies as it is an energetically costly behavior and may expose the animal to predation. However, erecting the body is beneficial for getting closer to a potential host and obtaining mechanosensory information when the host is within tangible reach. I thus postulate that after CO₂ exposure, ticks are more motivated to seek mechanoreceptive touch stimuli that enable them to attach to hosts. This idea is consistent with the observation here that after exposure to CO₂, ticks were more likely to grab onto a paintbrush that was brought into their immediate proximity. In contrast, animals not exposed to CO₂ crouched or did not react. While these observations were not quantified, they suggest that CO₂ activates and promotes mechanosensory host seeking in I. scapularis.

Mechanosensation in ticks is vastly unexplored, and to my knowledge, there are no examples from other animal species that CO₂ primes tactile stimulation. However, ticks belong to the class Arachnida along with spiders, scorpions, mites and harvestmen, many of which have highly developed mechanosensory systems for detecting prey and predators, as well as for proprioception and communication with potential mates (Barth, 2002). The exoskeleton of many arachnids is covered with mechanosensory touch hairs or sensilla. The cuticula and legs of I. scapularis are also covered with hair-like sensilla. It is reasonable to assume that many of these sensilla are mechanically sensitive and involved in controlling body and joint positions. However, I suspect that mechanosensory sensilla on the forelegs are also essential for host attachment. An electron microscopic examination of the foreleg tarsi of the cattle tick Rhipicephalus microplus (formerly Boophilus microplus) has shown that the morphology of sensilla surrounding the HO resembles that of mechanosensors (Waladde, 1977), supporting the present hypothesis.

So far, the only CO₂-sensitive structures identified in ticks have been found in the HO capsule of the tropical bont tick A. variegatum (Steullet and Guerin, 1992). Two of the seven capsule sensilla in A. variegatum are sensitive to changes in CO₂ concentration. CO₂ elicits opposite responses in the afferent neurons of the two sensilla. In one sensillum, neurons decrease their action potential firing rate as soon as CO₂ levels are about 0.002% above ambient, while neurons in the other sensillum are less sensitive and increase their firing rate at CO₂ levels above 0.1%. Apart from this study, no other examples of CO₂-sensitive sensilla in other tick species are available and it is unclear whether the HO of I. scapularis bears similarly CO₂-sensitive sensilla.

The present results demonstrate that I. scapularis still respond to CO₂ even when the HO is disabled (Figs 4 and 5). In fact, the CO₂ responses were only slightly affected by disabling the HO. Slightly fewer animals responded and the response onset was delayed in comparison to ticks with intact HO. This result was surprising and prompted extensive control experiments to rule out experimental artifacts. I doubled the number of animals tested and determined that the wax used to cover the forelegs is CO₂ impermeable (Fig. 4B). I also ruled out that the wax application itself might have affected the animals' behavior (Fig. S5). I even amputated the forelegs and the HO, and the animals still showed robust responses to CO₂. Taken together, these data thus demonstrate that the HO is not necessary for CO₂ detection and the subsequent initiation of host-seeking behavior.

However, this does not imply that the HO is not involved in CO₂ detection. The data indicate that the HO might be able to detect CO₂. When I examined the startle responses that occurred immediately after treatment onset, I found that they were not independent of the CO₂ concentration. Although ticks showed startle responses during air and CO₂ stimuli, the number of responding animals was highest at the lowest CO₂ concentration (1%; Fig. S2B). This suggests that the presence of air movement is recognized but that low levels of CO₂ enhance the startle response. The puff duration at 1% CO₂ was the shortest (0.58 s). The puff onset and offset may cause vibration of the air. Whether the increased number of startle responses at 1% CO₂ is mediated via high-frequency tuned mechanoreceptors or by CO₂-sensitive structures in the HO remains to be tested. However, the present data show that startle responses are mediated by the HO. Startle responses were almost absent when the HO was disabled (Fig. S4C), supporting the idea that one of the functions of the HO is to sense CO₂ plumes and their edges actively. This further implies that the HO of I. scapularis contains CO₂-sensitive afferent neurons that respond to low concentrations.

The main finding of the present study is that there must be a parallel CO₂-detection pathway that initiates the orientation phase of the host-seeking behavior, i.e. questing and faster walking. While I could not identify the CO₂ detectors that initiate the behavioral responses, I can rule out the Haller's and palpal organs (Figs 4 and 5). A possible candidate is the paired spiracle plate located on the ventrolateral surface of the tick body directly posterior of each hindleg coxa. The spiracle plates are the main respiratory organ in ticks and important for gas exchange (Soneshine and Roe, 2014). Regardless of the structure that mediates CO₂ detection, it is most likely conveyed via passive sensing. In comparison to active sensing where energy is invested to sample the environment, passive sensing is cost effective. However, the disadvantage of passive sensing is that it depends on the availability and quality of the stimulus. It typically has lower resolution and accuracy, and is often inadequate for determining target distance, speed and direction. This fits my observation that the behavioral responses were independent of CO₂ concentration, and it is consistent with the idea that there are two parallel CO₂ detection systems in ticks. The passive system detects the presence of CO₂ and initiates active sensing by foreleg waving to detect potential hosts in the area. The foreleg sensory structures, including the HO, then seek out host-emitted cues such as odors, body heat and mechanosensory touch stimuli. Further studies on tick sensory physiology and neuronal innervation and projections are needed to unravel the definitive role of these two sensory pathways.

I thank Wolfgang Stein (Illinois State University), Fernando Vonhoff (University of Maryland, Baltimore County) and Kim-Ann Saal (Universitätsmedizin Göttingen) for helpful discussions and comments on the manuscript, and Silvio Rizzoli (Universitätsmedizin Göttingen), Mark Frye (University of California, Los Angeles) and Wolfgang Stein for support and mentorship. Many thanks also to the Grass Foundation and 2021 Grass Fellows Bernardo Pinto, Duncan Leitch, Oscar Arenas Sabogal and Luis Bezares Calderon, as well as the 2021 Grass Director Melissa Coleman and Associate Director Laura Cocas and her postdoc Daniela Moura as well as the Trustees of the Foundation for support and constructive feedback.