香京julia种子在线播放

The thermoregulatory capacity of three species representing three orders was assessed during the 2018–2019 austral summer (2 December – 4 February) at the 76-km² Black Mountain Mine Conservation Area (29°18′S, 18°51′E) near Aggeneys, Northern Cape, South Africa. The Black Mountain Mine Conservation Area is situated within the Koa River Valley, which consists predominantly of red sand dunes with scattered shrubs (Rhigozum trichotomum) and large-seeded grasses (Brachiaria glomerata, Stipagrostis amabilis, Stipagrostis ciliata, and Stipagrostis brevifolia). In parts of the study site, the dunes are separated by unvegetated gravel plains. The climate is arid with mean annual precipitation of ∼100 mm and daily maximum T_air that ranged from 26.4 to 40.3°C during the study period (Kemp et al., 2020). The study species varied in M_b, as well as foraging behaviour, water dependence and activity phase (Table 1). We captured eight Namaqua sandgrouse (hereafter, sandgrouse), 10 spotted thick-knees (hereafter, thick-knees), and seven spotted eagle-owls (hereafter, eagle-owls). Namaqua sandgrouse were captured at a regularly frequented drinking spot using mist nets. Eagle-owls were captured at night opportunistically using a modified spring trap baited with two live mice. Thick-knees were captured at night using a hand-held net and torch.

Captured birds were placed in clean cloth bags and transported (approximately 45-min trip by vehicle) to an animal holding room where they were kept indoors in individual cages (0.8 m³). Individuals were kept in cages for no more than 12 h and provided with an ad libitum supply of water. In the case of the eagle-owls and thick-knees, experimental procedures took place the morning following capture, ≥10 h after birds last fed. In the case of sandgrouse, the time between capture and being placed in respirometry chambers was ∼ 3 h.

A temperature-sensitive passive integrated transponder (PIT) tag (BioTherm13, Biomark, Boise, ID, United States) injected into each individual’s abdominal cavity measured T_b. We calibrated these tags in a circulating water bath (model F34; Julabo, Seelbach, Germany) over a 35–50°C range against a thermistor probe (TC-100; Sable Systems, Las Vegas, NV, United States). We placed PIT-tagged individuals in a metabolic chamber placed next to an antenna connected to a transceiver system (HPR+; Biomark). The PIT tags we used for this study were from the same batch as the 23 tags calibrated by Freeman et al. (2020), who found that measured values deviated from actual values by 0.28 ± 0.23°C. We inserted a thermistor probe (TC-100; Sable Systems) sealed with a rubber grommet through the approximate centre of the lid of each metabolic chamber to measure T_air during the gas exchange measurements.

We measured EWL and carbon dioxide production (V.CO2) using an open flow-through respirometry system. The metabolic chamber consisted of either a 12.8-L (sandgrouse) or 60-L (eagle owl and thick-knee) air-tight plastic container (previously shown to not adsorb water vapour; Whitfield et al., 2015). Plastic mesh platforms were placed in the metabolic chambers to ensure individuals were elevated at least ∼10 cm above a ∼1 cm layer of mineral oil. These plastic chambers were situated within a temperature-controlled custom-built aluminium box (∼640 L). The T_air within the chamber was controlled via a Peltier device (AC-162 Thermoelectric Air Cooler; TE Technology, Traverse City, MI, United States) connected to a digital controller (TC-36-25-RS485 Temperature Controller; TE Technology) and mounted to the side of the box.

Atmospheric air was supplied by an oil-free compressor and passed through a membrane dryer (Champion CMD3 air dryer and filter; Champion Pneumatic, Princeton, IL, United States) to scrub water vapour before being split into baseline and experimental channels. A needle valve (Swagelok, Solon, OH, United States) regulated flow rate in the baseline channel and a mass flow controller (MC50 SLPM, Alicat Scientific Inc., Tucson, AZ, United States), calibrated using a soap-bubble flow meter (Gilibrator-2; Sensidyne, St. Petersburg, FL, United States), regulated flow rates in the experimental channel. To maximise air mixing, and minimise any potential convective cooling at higher flow rates, we positioned the air inlet near the top of the metabolic chamber with an elbow joint facing upward and the air outlet below the mesh platform. Flow rates (13–65 L min^–1) were adjusted to maintain humidity at a dewpoint <−5°C within the chamber, depending on T_air, M_b and individual behaviour, while still allowing for the accurate measurements of differences in water vapour and CO₂ between incurrent and excurrent air. A respirometry multiplexer (model MUX3-110118M, Sable Systems) in manual mode and an SS-3 Subsampler (Sable Systems) subsampled excurrent air from the baseline and chamber air. Subsampled air was pulled through a CO₂/H₂O analyser (model LI-840A; LI-COR, Lincoln, NE, United States). The CO₂/H₂O analyser was regularly zeroed using nitrogen, and spanned for CO₂ using a certified calibration gas (CO₂ concentration = 2,000 ppm; Afrox, South Africa). We regularly zeroed the H₂O sensor of the LI-840A using nitrogen and spanned it using air with a known dewpoint produced by a dewpoint generator (DG-4; Sable Systems). The system’s tubing was Bev-A-Line IV tubing (Thermoplastic Processes Inc.). An analogue-digital converter (model UI-3; Sable Systems) digitised voltage outputs from the thermistor probes and analysers, which were recorded every 5 s using Expedata software (Sable Systems).

Following Czenze et al. (2020) we placed individuals in chambers prior to measurements to habituate. Before measurements commenced, individuals were held without food for at least 1 h in the metabolic chamber, and together with the preceding period of fasted captivity ensured birds were likely postabsorptive when data collection started (Karasov, 1990).

Measurements took place during the day and individuals were exposed to a stepped series of progressively higher T_air values over which V.CO2 and EWL were recorded. We started measurements with a baseline air subsample until water and CO₂ readings were stable (∼5 min). Chamber excurrent air was then subsampled once T_air had stabilised at the target value. We recorded data at this T_air until V.CO2 and H₂O traces were stable for at least 10 min, with individuals spending approximately 20–25 min at the target T_air before we switched back to baseline air for another 5 min. Trials began at 28°C with 4°C increases until 40°C and then 2°C increases until a maximum of 60°C (the maximum T_air our thermistors could accurately measure). We monitored individuals continuously using an infrared video camera. Individuals were removed from the chamber when they reached their thermal endpoints [i.e., loss of coordination or balance, sudden and rapid decrease in EWL or resting metabolic rate (RMR), rapid and uncontrolled increase in T_b to values >45°C] or displayed sustained escape behaviours like agitated jumping. In the case of the sandgrouse, measurements were terminated at T_air slightly above 60°C. Once an individual was removed from the chamber we quickly dabbed its underbelly with ethanol to facilitate rapid cooling and held it under chilled air produced by an air conditioner. Once the individual’s T_b was stable (40–42°C), we offered it water using a syringe, and placed it back in its cage to rest, with ad libitum water available. The bird was later released at the site of capture. This experimental protocol has been used previously on birds and bats and individuals recaptured several days or weeks post-release showed no adverse effects (Kemp and McKechnie, 2019; Czenze et al., in press).

We used the R package segmented (Muggeo, 2008) to perform a broken stick regression analysis in R 3.5.2 (R Core Team, 2018) to identify inflection points for EWL, RMR, T_b, and EHL/MHP. To determine whether a broken-stick regression fit the data better than a simple linear model, we compared broken-stick models against generalised mixed-effect models created in R package nlme (Pinheiro et al., 2009) using ANOVA. If we retained a broken stick regression and significant inflections occurred, we analysed data below and above inflection points separately when estimating slopes for the relationships of EWL, RMR, EHL/MHP, and T_b as functions of T_air. We created general linear mixed-effect models for each species using the R package nlme to test for an effect of T_air on each response variable above inflection points. We accounted for the repeated measures design of our study (i.e., measurements at multiple T_air values per individual) by including individual identity as a random factor in all analyses. Initial models contained T_air, M_b, and the T_air:M_b interaction and model selection was performed using the “dredge” function in the MuMIn package (Bartoń, 2013). We selected the model with the highest rank among competing models using Akaike weights and Akaike information criterion values corrected for small sample size (AICc) (Burnham and Anderson, 2002). Body mass did not emerge as an important predictor for any of the response variables of any species. We assessed significance at α < 0.05 and values are presented as mean ± SD.

Normothermic T_b at thermoneutrality varied by ∼1.4°C among species, from 39.1 ± 0.4 (°C) in eagle-owls to 40.5 ± 0.4 (°C) in thick-knees (Figure 1 and Table 2). All three species showed significant increases in T_b above normothermic levels when exposed to higher T_air (Figure 1). In thick-knees and eagle-owls, significant inflections above which T_b increased rapidly occurred at T_air = 40.7°C and T_air = 50.9°C, respectively (Table 2 and Figure 1). At T_air values above these inflection points, T_b increased significantly (t = 16.94, P < 0.001 and t = 7.36, P < 0.001, respectively) and in a linear fashion in thick-knees and eagle-owls. In sandgrouse, on the other hand, no inflection was evident and a linear model provided the best fit for the relationship between T_b and T_air ≥ 32.5°C [the value we assumed to represent this species’ lower critical limit of thermoneutrality (Figure 2)]. Over this T_air range, T_b increased significantly (t = 18.53, P < 0.001) to a maximum of 43.2 ± 0.8°C (n = 7). Mean maximum T_b values at very high T_air varied among the three species from 43.2°C to 44.3°C (Table 2). Heat tolerance limits (HTL, i.e., maximum T_air tolerated before thermal endpoints reached) varied from 56°C in thick-knees to 60°C in eagle-owls and Namaqua sandgrouse (Table 2), although HTL for the latter species may be even higher in some individuals.

Minimum thermoneutral RMR values varied from 0.83 W in Namaqua sandgrouse to 2.38 W in the eagle-owl (Table 2). Namaqua sandgrouse did not show a significant inflection in RMR, which did not vary significantly with T_air (Figure 2). In contrast, significant inflections in RMR were evident for both thick-knees and eagle owls (Table 2), above which linear increases in RMR were significant (t = 9.15, P < 0.001 and t = 4.1, P < 0.001, respectively; Figure 2). These inflection T_air values, which correspond to the upper critical limits of thermoneutrality (T_uc), were T_air = 47.3°C in thick-knees and T_air = 49.2°C in eagle-owls. Maximum RMR at the highest T_air achieved was equivalent to 137%, 196%, and 276% of minimum thermoneutral values in Namaqua sandgrouse, eagle-owls and thick-knees, respectively (Figure 2). Mass-specific maximum RMR in thick-knees was ∼85% higher compared to sandgrouse or eagle-owls (Table 3).

Gular flutter commenced at T_air between 36.7°C and 44.5°C and T_b between 39.4°C and 42.4°C (Table 2). Minimum EWL varied nearly 2-fold from 1.01 g h^–1 in Namaqua sandgrouse to 1.83 g h^–1 in thick-knees (Table 2 and Figure 3). All species showed clear inflection points above which EWL increased rapidly and linearly; these varied from T_air = 33.9°C in Namaqua sandgrouse to T_air = 42.8°C in thick-knees (Table 1 and Figure 3). Increases in EWL were significant for sandgrouse (t = 19.94, P < 0.001), thick knees (t = 13.67, P < 0.001), and eagle owls (t = 19.59, P < 0.01). Maximum rates of EWL ranged from 7.48 g h^–1 in Namaqua sandgrouse to 23.52 g h^–1 in thick-knees, with the highest mass-specific values in thick-knees (Table 3). The fractional increase in EWL (i.e., the ratio of maximum EWL to minimum EWL) ranged from ∼7.41 for Namaqua sandgrouse to ∼12.85 for thick-knees (Table 2).

The relationships between evaporative heat loss (EHL)/metabolic heat production (MHP) and T_air–T_b gradients (Figure 4), were characterised by significant inflections for all species (Table 2). Above these inflection points, EHL/MHP increased linearly with increasing T_air–T_b, with relationships significant for sandgrouse (t = 15.63, P < 0.001), thick knees (t = 11.62, P < 0.001) and eagle owls (t = 11.30, P < 0.001). Maximum evaporative cooling efficiency (i.e., maximum EHL/MHP) ranged from 2.75 for eagle-owls to 5.49 in Namaqua sandgrouse (Figure 4). However, these EHL/MHP values were calculated assuming RER = 0.71 (i.e., lipid catabolism), and could be higher if individuals used carbohydrates (i.e., RER = 1.00). In this case, maximum EHL for eagle-owls and Namaqua sandgrouse would shift upward to 3.66 and 7.30, respectively. Visual inspection of the data in Figure 4, however, does not support this possibility. Theoretically, evaporative heat loss and metabolic heat production should be equal (i.e., EHL/MHP = 1.0) when T_air = T_b. Currently EHL/MHP = 1.0 occurs at approximate T_air – T_b values of −1.5, −2.0, and −4.5°C for thick-knees, eagle-owls, and Namaqua sandgrouse, respectively. Recalculating EHL/MHP using RER > 0.71 would shift these values even further below T_air – T_b = 0, suggesting our assumption that RER = 0.71 is likely correct.

Relationships between T_b and T_air conformed to typical avian patterns in thick-knees and eagle-owls, although the inflection T_air for eagle-owls is well above the range of 30–40°C typical of most species investigated to date (McKechnie et al., 2017; McWhorter et al., 2018; Smit et al., 2018; Czenze et al., 2020). The eagle-owl inflection is also substantially higher than those reported for two smaller owls; T_air = 37.3°C in 101-g western screech-owls (Megascops kennicottii) and T_air < 30°C in 40-g elf owls (Micrathene whitneyi; Talbot et al., 2018). In contrast to the thick-knees and eagle-owls, no significant inflection occurred in the relationship between T_b and T_air in Namaqua sandgrouse, a pattern qualitatively similar to those reported for P. namaqua and double-banded sandgrouse (P. bicinctus) held in outdoor enclosures during summer in the Namib Desert (Thomas and Maclean, 1981), as well as black-bellied sandgrouse (P. orientalis) at T_air between 10 and 45°C (Hinsley et al., 1993).

Maximum T_b at thermal endpoints is within the range documented for non-passerines (McKechnie et al., 2021a), with values >44°C for thick-knees and eagle-owls near the upper end of this range. The heat tolerance limits of all three species are higher than those of most non-passerines and all passerines, with the values of 58°C and 60°C in eagle-owls and Namaqua sandgrouse, respectively, comparable to those of some caprimulgids and columbids, the avian taxa with the highest documented heat tolerances (Marder and Arieli, 1988; O’Connor et al., 2017; Talbot et al., 2017).

The linear increases of EWL with T_air > T_b also followed patterns typical of birds, although evaporative scope (i.e., maximum EWL/minimum EWL) was nearly twice as high in thick-knees (12.9) compared to Namaqua sandgrouse (7.4) or eagle-owls (7.8). Moreover, maximum rates of EWL for thick-knees (23.5 g h^–1 at T_air = 56°C) exceed by a substantial margin the highest rates of EWL reported for the much larger (∼3-fold) MacQueen’s bustard (Tieleman et al., 2002). Evaporative water loss rates of thick-knees were equivalent to 5.4% M_b h^–1, higher-than-expected fractional values more similar to those typically seen in small passerines (Wolf and Walsberg, 1996; McKechnie et al., 2017; Czenze et al., 2020).

Increases in RMR associated with evaporative heat dissipation at high T_air varied substantially. In Namaqua sandgrouse, maximum RMR at T_air = 60°C was only 37% higher than minimum thermoneutral values, a pattern similar to those observed in some caprimulgids (O’Connor et al., 2017), columbids (McKechnie et al., 2016a,b) and at T_air ≤ 55°C in MacQueen’s bustard (Tieleman et al., 2002). Among other sandgrouse species, no daytime T_uc was evident in P. orientalis nor P. bicinctus at T_air values below ∼43°C (Hinsley, 1992; Hinsley et al., 1993), suggesting measurements may not have reached T_air high enough to elicit a T_uc in either of these species. The very high T_uc of eagle-owls (T_air = 49°C) is similar to the T_uc = 46.4°C reported for western screech owls (Talbot et al., 2018) and other taxa, including lilac-breasted roller (Coracias caudatus: T_uc = 47.5°C; Smit et al., 2018) and several caprimulgids and columbids (Smith et al., 2015; McKechnie et al., 2016a; Talbot et al., 2017).

With the exception of maximum EHL/MHP = 3.4 in lilac-breasted rollers, the values we report here are the highest yet documented among birds other than caprimulgids and columbids (reviewed by McKechnie et al., 2021a). In particular, maximum EHL/MHP = 5.5 in Namaqua sandgrouse is the highest avian value measured to date, exceeding values of 4.7 and 5.2 previously reported for Namaqua doves (Oena capensis; McKechnie et al., 2016a) and Rufous-cheeked nightjars (Caprimulgus rufigena; O’Connor et al., 2017), respectively. We are not aware of published data on maximum EHL/MHP in Charadriiformes, but our value of 3.2 for thick-knees raises the possibility that the capacity for evaporative cooling may also be pronounced in this order (e.g., Grant, 1982).

The data currently available suggest patterns of thermoregulation in the heat are phylogenetically conserved within avian orders. The large differences in patterns of thermoregulation at very high T_air between P. namaqua and the congeneric Burchell’s sandgrouse (P. burchelli, McKechnie et al., 2016b) were, therefore, surprising. Whereas RMR increased sharply to values equivalent to ∼250% of minimum thermoneutral levels in P. burchelli, the RMR of P. namaqua remained within 37% of minimal levels (Figure 5). Consequently, maximum EHL/MHP was substantially higher in P. namaqua (5.5 at T_air = 60°C) compared to P. burchelli (2.0–2.7 at T_air = 56°C). The rapid increases in RMR and modest maximum EHL/MHP of P. burchelli bear a strong resemblance to the pattern typical of passerines, which lack gular flutter or rapid cutaneous water loss and rely on panting for evaporative heat dissipation. In contrast, the corresponding patterns for P. namaqua resemble the extremely efficient cooling characteristic of columbids (Smith et al., 2015; McKechnie et al., 2016a) and caprimulgids (O’Connor et al., 2017; Talbot et al., 2017). The sandgrouse provide the first indication that fundamental differences in evaporative cooling processes can occur within an order, or even a single genus.

We speculate the differences between these two species reflect the more arid distribution of P. namaqua, which likely experiences stronger selection to minimise energy and water requirements. The distribution of P. burchelli is centred on the Kalahari basin, spanning arid savanna in the south to mopane (Colophospermum mopane) woodlands in the north. Much of the distribution of P. namaqua, on the other hand, is more arid, including the Nama Karoo and hyperarid Namib Desert. With mean annual precipitation of ∼100 mm y^–1, our present study site is considerably more arid than any part of the range of P. burchelli (Lloyd, 2005). The methods used to quantify thermoregulatory responses in both species were identical, but we cannot rule out possibilities such as a greater stress response in P. burchelli compared to P. namaqua. However, in this scenario we would expect higher T_b under thermoneutral conditions in P. burchelli on account of stress-induced hyperthermia (Cabanac and Briese, 1992; Cabanac and Aizawa, 2000; Nord and Folkow, 2019). The similarity of both normothermic T_b (P. burchelli: 39.0°C and P. namaqua: 39.4°C) and maximum T_b (P. burchelli: 43.6°C and P. namaqua: 43.2°C) argues against the notion that the divergent responses of these congeners reflect an artefact of stress responses to the experimental conditions.

The high daytime environmental temperatures nocturnal birds endure in hot habitats, combined with their lack of drinking during the day, led O’Connor et al. (2018) to hypothesise that nocturnal species have evolved more economical evaporative cooling and reduced EWL to minimise dehydration risk during the day. An analysis of data for 32 diurnal and 7 nocturnal species (two owls, four caprimulgids and one owlet-nightjar) provided weak support: slopes of EWL between T_air = 40 and 46°C were significantly lower in nocturnal species, but neither log₁₀ EWL nor EHL/MHP at T_air = 46°C differed significantly with activity period (O’Connor et al., 2018). Our data for thick-knees and eagle-owls generally support O’Connor et al.’s (2018) hypothesised link between nocturnality and more economical evaporative cooling, and specifically the predictions of lower EWL (Figure 6A) and higher EHL/MHP (Figure 6B) in nocturnal species, as well as these authors’ contention that their results reflected a small sample size for nocturnal species.

The data we present here reveal that representatives of three non-passerine orders can achieve rapid evaporative heat dissipation and tolerance of T_air = 55–60°C, similar to the performance of smaller species in highly heat tolerant taxa such as Caprimulgiformes and Columbiformes. Most species in which heat tolerance has been investigated so far are smaller than 100 g, and interspecific variation in interactions between evaporative heat loss and metabolic heat production at T_air > 45°C in larger species has received little attention (Table 4). The contrast between MacQueen’s bustards (Tieleman et al., 2002) and brown-necked ravens (Marder, 1973) is particularly striking (Table 4). Despite being twice as large, bustards’ EWL and RMR were equivalent to <50% of corresponding values in ravens, whereas EHL/MHP was similar and T_b was 3.2°C higher in ravens (Table 4). Although the relative humidity experienced by ravens (<30%, Marder, 1973) may have been higher than for bustards (<22%, Tieleman et al., 2002), we suspect these differences primarily reflect the ravens’ reliance on panting and the constraints imposed by this relatively inefficient avenue of evaporative cooling (McKechnie et al., 2021a). Spotted eagle-owls (present study) maintained the lowest T_b at high T_air of any of the six species, via a combination of intermediate rates of EWL and low RMR (Table 4).

The pronounced interspecific variation in EWL at T_air = 50°C reveals that vulnerability to lethal dehydration during extreme heat events differs considerably among the six species listed in Table 4. Assuming birds can tolerate losing water equivalent to 15% of M_b [following Albright et al. (2017) and others], survival times at T_air = 50°C vary from 4.53 h (spotted thick-knee) and 5.13 h (brown-necked raven) to 8.88 h (spotted eagle-owl) and 21.61 h (MacQueen’s bustard). That estimated survival times for these species are >4 h at this extremely high T_air emphasizes the effect of M_b on the risk of lethal dehydration during extreme heat events; larger birds are at considerably lower risk of mortality via dehydration tolerance limits being exceeded (McKechnie and Wolf, 2010).