INTRODUCTION

Top Abstract Introduction Methods Results Discussion Appendix References

MANY SPECIES of birds and mammals forage underwater. Diving behavior, which often exhibits a distinct daily rhythmicity (4, 14, 21, 24, 33, 34), typically consists of bouts of aerobic dives (31, 35). These dives tend to be short compared with maximal breath-hold times, and they are separated by similarly brief intervals at the water surface (24, 31).

Current models of the physiological regulation of dive patterns are either explicitly based on, or implicitly incorporate, aspects of the "aerobic dive limit" concept (22). Although not discounting the potential for short-term modification of dive patterns through anaerobic mechanisms (9), the aerobic dive limit hypothesis predicts that, during routine aerobic diving behavior, dive times are directly or indirectly determined by the quantity of O₂ stored in the lungs, blood, and muscle and the rate at which the O₂ is consumed (22). If O₂ or a correlated variable serves as a signal to terminate a dive (i.e., to resume lung ventilation), then the respiratory sensitivity to hypoxemia or a correlated variable is an integral component of this model. It has been suggested that diving species have a relatively low respiratory responsiveness to asphyxia (6), although supporting evidence is scarce. It is assumed that this adaptation would enable diving species to utilize a greater fraction of their O₂ store and therefore to remain submerged for longer.

In this study, it was hypothesized that the circadian organization of avian diving behavior may reflect an underlying circadian oscillation in respiratory chemosensitivity. Day-night differences in respiratory responses to hypoxia and hypercapnia have been demonstrated in rodents (26, 27, 30) and humans (28), but no information is available for birds or diving species. To test this hypothesis, the circadian rhythm in diving behavior was monitored in captive canvasback ducks, and the ventilatory response to progressive asphyxia was measured at times corresponding to maximal and minimal dive activities.

	METHODS

Top Abstract Introduction Methods Results Discussion Appendix References

Ten adult canvasback ducks (Aythya valisineria; 7 females, 3 males; body mass range 900-1,200 g) were housed in an indoor diving tank under a 12:12-h light-dark photoperiod (lights on at 6:30 AM EST). The ducks were trained to dive for food that was delivered to the floor of the tank by a semiautomatic feeder. This feeder, which was based on the design used by Bevan et al. (2), provided the ducks with unlimited access to food. The ducks were trained to activate a submerged paddle switch to deliver 1.5-2 g of food pellets. Ducks used this self-feeding system for several weeks and were observed to maintain normal body weight before recordings were initiated.

Animal preparation. Six of the ducks underwent surgery a minimum of 1 wk before experiments. All procedures were performed under general anesthesia (halothane, 1-2.5% in a 50% air-50% O₂ mixture) using sterile materials. A temperature-sensitive radiotransmitter was implanted into the abdomen through a midline incision and sutured to the muscle of the abdominal wall. Care was taken to avoid damaging the abdominal plumage so that the ducks could resume diving behavior within 1 day of recovery. The ducks were then given a topical antibiotic (Cicatrin) and a broad-spectrum antibiotic (Penlong XL, Rogar/STB, 0.3 ml/kg im). The ducks were kept in postsurgical isolation for 2 h until full recovery from anesthesia and then returned to the diving tank. The six animals with radiotransmitters were used for measurement of circadian rhythms in dive duration and for plethysmographic analysis of ventilatory sensitivity to progressive asphyxia.

Diving behavior. An indirect measure of the temporal organization of diving behavior of the group of 10 ducks was obtained by recording activations of the feeding device. The circadian rhythm was first measured during a 2-wk control period during which the ducks were left undisturbed (with the exception of an experimenter filling the food hopper once every 4 days and the cleaning the tank once per week). Measurements continued during subsequent experiments to ensure that disturbance and handling did not disrupt the diving rhythm.

Measurement of ventilation. A closed-circuit plethysmography system was used to measure resting lung ventilation in six animals (Fig. 1). The system operates on the principle that air is warmed and humidified during inspiration causing an expansion in volume, which in turn results in an increase in pressure within the fixed-volume animal chamber (15). The ventilatory pressure waveform was monitored using a differential pressure transducer (model DP45-14, Validyne Engineering), which measured the difference in pressure between the animal chamber and an identical reference chamber. The ventilatory pressure signal was calibrated using an animal ventilator (Columbus Instruments). T_b was continuously recorded using implantable radiotransmitters (Mini-mitter; TM-DISC or VM-FH-DISC). The transmitters weighed ~4.9 g when fully encapsulated, they had a response time constant of 77 s and a resolution of 0.1°C. Nasal temperature was also continuously recorded using a thermocouple probe (Physitemp Instruments, EXT-6) taped to the bill at the opening of the nares. Chamber temperature and humidity were recorded periodically from a small thermometer and hygrometer located in the animal chamber.

View larger version (35K): [in this window] [in a new window]   Fig. 1.   Schematic diagram of closed-circuit plethysmograph. Dotted lines represent electrical connections. Bold lines represent primary recirculating airflow. Fine lines represent airflow for gas analysis and O2 supply. 1, Three-way stopcock used to include or exclude KOH solution (CO2 scrub) in primary air circulation; 2, 3-way stopcock used to enable CO2 analyzer to sample gas before or after KOH solution.

(1)

Experimental protocol. At the start of each experiment, the animals were allowed to rest quietly in the chamber until all variables had reached an apparent steady state. During this period, which usually lasted ~30 min, O₂ was bled into the plethysmograph at a flow rate sufficient to maintain the O₂ concentration between 20.98 and 21.02%. CO₂ concentration was kept at 0.05% by bubbling the gas through concentrated KOH solution. A recording of CO₂ production and lung ventilation was made for 5 min; then the duck was exposed to progressive asphyxia during a period of rebreathing. To accomplish this, the O₂ supply was turned off, and the CO₂ scrubbers were temporarily bypassed. Lung ventilation, gas exchange, and temperature were recorded continuously during the rebreathing period. The experiment was terminated when CO₂ concentration reached 6%. Each duck was used three times at 9 AM and three times at 9 PM, and at least 60 h separated consecutive experiments.

Analysis of plethysmography data. Tidal volume (VT, ml BTPS) was calculated as follows (18)

(2)

where

(3)

and

(4)

where Pm is the observed ventilatory pressure deflection; Pcal is the pressure deflection due to cyclic injection and withdrawal of a known volume, Vcal; Ga represents the ratio of tidal volume at respiratory air sac conditions to the small increase in volume that occurs when the tidal gas expands from its volume at chamber conditions; Gn is analogous to Ga but represents the volume change from air sac to nasal temperature, T_n (K); tI is the inspiratory time (s); t_tot is the total breath time (s); PB is the barometric pressure (mmHg); Pc_H2O is the water vapor pressure of gas in the animal chamber (mmHg); Pn_H2O is the saturated water vapor pressure of gas at T_n (mmHg); Pa_H2O is the saturated water vapor pressure of gas in the respiratory air sacs (mmHg); T_b is abdominal air sac temperature (K); and T_c is the temperature (K) of gas in the animal chamber.

(5)

(6)

(7)

	RESULTS

Top Abstract Introduction Methods Results Discussion Appendix References

Diving activity exhibited a daily rhythm (Fig. 2). In 10 undisturbed ducks, the underwater feeder was activated a total of 8,273 times in 12 days. Seventy-six percent of the dives occurred during darkness and 24% in the light. The lights-on and lights-off transitions were followed by abrupt increases in diving activity. The increase in diving activity at lights-on was transient (1-2 h), whereas that at lights-off was maintained for ~8 h (Fig. 2). Feeding dives were least frequently performed between ~8 and 11 AM, and the greatest number occurred between ~9 and 11 PM (Fig. 2). In six of the ducks, dive durations, surface intervals, and T_b during a diving bout and during nondiving intervals were analyzed over a 2-h time bin in these two time periods (Table 1). Dive durations were 37% greater at 9-11 PM than at 9-11 AM (P = 0.01). There was no significant difference in surface intervals at these two time intervals. T_b during diving bouts and nondiving periods was significantly higher at 9-11 AM than at 9-11 PM (P = 0.03). In addition, T_b was significantly higher during diving bouts than during nondiving periods at 9-11 AM and 9-11 PM (P = 0.003).

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Total number of feeding dives performed per hour by 10 ducks over 12 days. Lights came on at 6:30 AM and went out at 6:30 PM as indicated by horizontal bar.

Respiratory variables measured during steady-state air breathing are presented in Table 2. None of the variables differed significantly between 9 AM and 9 PM. Respiratory variables measured during progressive asphyxia are presented in Table 3. There were no significant differences in metabolic rate (O₂, CO₂, RER), breathing pattern (f_r, tI/t_tot), or T_b between 9 AM and 9 PM in the asphyxic state. There were, however, significant differences in T_b between steady-state air breathing and asphyxia (P = 0.0002) at 9 PM but not at 9 AM. T_b decreased progressively during asphyxia to minima of 40.6 ± 0.2 and 40.8 ± 0.2°C at 9 PM and 9 AM, respectively.

During progressive asphyxia, inspired ventilation (I) increased exponentially with both increasing CO₂ concentration (Fig. 3) and decreasing O₂ concentration. The ventilatory response was analyzed as a function of CO₂ concentration, and an analysis of the response as a function of O₂ concentration yields the same qualitative result because RER was invariant in each experiment.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Day and night ventilatory responses to progressive asphyxia in a representative animal. , 9 AM data; , 9 PM data.

VT, I, VT/tI, and I/CO₂ were analyzed as a function of inspired CO₂ concentration (%) using an exponential model of the following type: y = a · e^b[CO₂], where y is a respiratory variable. The coefficient a represents the magnitude of y under air-breathing conditions, and the exponent b is interpreted as a measure of the sensitivity of y to progressive hypercapnic hypoxia. The exponent b was derived by least-squares linear regression of lny against CO₂ concentration ([CO₂]). The resulting r² values were slightly, but consistently, higher than for linear regressions of the untransformed data. Furthermore, examination of plots of residuals vs. predicted values showed that untransformed data were heteroscedastic and exhibited a curved relationship to the regression. These undesirable characteristics were corrected by logarithmic transformation, indicating that the exponential model provided a better description of the data than a linear model. There were significant day-night differences (P < 0.05) in the value of the exponential rate constant b, i.e., in the slopes of lnVT/[CO₂], lnI/[CO₂], ln(I/CO₂)/[CO₂], and ln(VT/tI)/[CO₂] (see Table 3).

A least-squares regression of f_r against [CO₂] yielded a nonsignificant regression coefficient (P = 0.527), indicating that f_r did not change during progressive asphyxia. Therefore an average value was calculated over the entire rebreathing period and is presented in Table 3.

There were no statistically significant differences in T_n between 9 AM and 9 PM in air or during asphyxia. In addition, T_n did not differ significantly between air and asphyxia at either time of day. The ducks often began to breathe through the open mouth when inspired CO₂ rose to ~6%. When this occurred, T_n increased instantaneously by ~4°C, and experiments were terminated.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion Appendix References

This study, which to our knowledge represents the first continuous long-term recording of duck diving patterns, has confirmed that ducks exhibit a circadian rhythm in diving behavior. Dives were more frequent and of longer duration at night. Furthermore, the plethysmographic data support the hypothesis that the ventilatory response to asphyxia is lower at night, when the ducks are diving the most, than in the morning, when they dive the least.

All animals responded to progressive asphyxia with a nonlinear increase in ventilation and no change in f_r. This corroborates other studies in birds in which hypercapnic hypoxia produced a nonlinear ventilatory response (12), with the increase in ventilation due mainly to a rise in VT with small changes in f_r (7, 12). Inhalation of CO₂ is known to evoke an increase in ventilation in unanesthetized birds (20, 32), although the effect on breathing pattern can be quite variable due, at least in part, to the inhibitory effect of CO₂ on the intrapulmonary chemoreceptors (25). In general, the avian hypercapnic ventilatory response consists of an increase in VT with small increases or decreases in f_r (1, 8, 20, 32). In addition, birds have been shown to respond to a decrease in ambient O₂ with an increase in ventilation, produced by increases in both VT and f_r (3, 5, 7, 8, 12, 19, 20). Comparisons of the ventilatory response to hypoxia, hypercapnia, and hypoxic hypercapnia indicate that effects of CO₂ and O₂ are additive in birds (20), although the data presented by Colby et al. (12) suggest that there may be a multiplicative interaction in adult bank swallows.

In the present study, it was found that simultaneous progressive hypoxia and progressive hypercapnia (asphyxia) evoked a greater ventilatory response in the day than at night. This difference in I, which was entirely due to a difference in VT, was manifest as a difference in the exponential rate constant b. Recent studies in humans (28) and rats (26, 27, 30) have provided evidence for the existence of a circadian rhythm in mammalian respiratory chemosensitivity, but the mechanisms of interaction of the circadian timing system and the respiratory control system are unknown at present. In rodents, it appears that the interaction could be mediated indirectly via circadian modulation of metabolic rate (26, 27, 30), but this was not the case in ducks. The exponential rate constant describing the relationship between I/CO₂ and inspired CO₂ concentration differed between 9 AM and 9 PM (Table 3). Thus the metabolism-specific ventilatory response to hypoxic hypercapnia was greater during the day than during the night.

The question as to whether diving mammals and birds have a reduced respiratory sensitivity to CO₂ is unresolved. Seals appear to have a lower sensitivity to CO₂ than humans (13, 29), but it is not established whether diving birds are less sensitive than nondiving birds (1, 7). It has been suggested that a chemoreceptor system with low sensitivity to CO₂ would allow a diving animal to remain submerged for extended periods of time (6, 13) because the drive to resurface would develop more slowly. In support of this idea, we observed a circadian difference in dive durations in the ducks in the present study, with longer dives occurring at night when respiratory sensitivity was low. Previous studies that did not convincingly demonstrate a lower chemosensitivity in aquatic birds (1, 5, 7, 25) may have been confounded by inappropriate timing of the tests. It is likely that experiments were done during the day when the respiratory system was most sensitive to chemical stimuli.

The correlation between high chemosensitivity and reduced dive times raises the question as to whether the respiratory chemosensory system plays a direct role in the control of dive times. If so, what kind of role is it? The aerobic dive limit concept is descriptive but not explanatory. That is, natural dive times of freely diving animals often fall within the predicted limit, but this does not tell us anything about whether O₂ or any other correlated variables actually act to signal the animal to resurface to breathe. A threshold mechanism based on O₂ content would implicate the respiratory chemosensory system, but such a scheme is likely to be too simplistic given what is known about the numerous mechanisms involved in respiratory control (11, 16).

We suggest an alternative hypothesis: that routine diving patterns are determined by the temporal characteristics of a sustained instability in the respiratory control system. The chemosensory system can be viewed as a damped feedback control loop with delay and variable feedforward drive (11, 16). Many diving vertebrates precede a series of dives with a short period of hyperventilation (6). This destabilizing maneuver is immediately followed by a step increase in metabolic rate and, through the reduction in cardiac output, a simultaneous lengthening of circulation time. Submersion of the head in water may also cause a step decrease in the central drive to breathe. The synchronized oscillations in metabolic rate (O₂, CO₂, and metabolic acid production), circulation delay, and perhaps central feedforward drive could conceivably give rise to periodic oscillations in respiratory output (11). This respiratory instability model predicts that the behavior of freely diving animals would become entrained to the natural frequency of oscillations in respiratory drive. The natural frequency of such oscillations is predicted to vary with such factors as the size of the O₂ and CO₂ stores, the buffering capacity of the body fluids, the intensity of the diving bradycardia, and the metabolic power input during underwater locomotion. In a feedback control system, for any given set of the above conditions, the amplitude and frequency of the oscillations would also be dependent on the gain (sensitivity) of the chemosensory system (11). Specifically, a reduction in the gain would tend to cause an increase in the period of the oscillation (i.e., increased dive times).

Some species of diving animals (e.g., whales, seals, penguins) may not show a circadian rhythm in respiratory chemosensitivity, because most of their activity budget involves diving, at least during foraging trips, which can last several weeks at a time (23, 24). Furthermore, it is likely that the control of dive times in ectothermic divers (e.g., turtles, sea snakes, crocodilians, amphibians, and air-breathing fishes) is more complicated than the proposed respiratory instability model would imply. The dependence of metabolism on such factors as water temperature and O₂ tension, the increased capacity for anaerobic metabolism, the capacity for cardiac shunting and cutaneous gas exchange, and the interdependence of respiration and buoyancy all suggest that the simple model suggested above may not represent a complete or adequate description of the respiratory control process in diving ectotherms. Nevertheless, daily rhythms in dive depths and duration have been recorded in some species (e.g., leatherback sea turtle; Ref. 17), and it would be interesting to know whether this behavior is correlated with changes in respiratory chemosensitivity.

Further research is required to determine whether observed daily diving rhythms represent endogenous circadian rhythms or whether they represent a direct effect ("masking") of rhythmic stimuli, such as the light-dark cycle or food availability. For example, many diving birds and mammals exhibit a daily pattern in both the time of day that they dive and in the depth and/or duration of those dives (4, 14, 21, 23, 24, 33, 34). However, this daily pattern is often correlated with daily changes in the abundance and distribution of their prey and may not represent an endogenous circadian rhythm. This confounding variable is not likely to apply to canvasback ducks, however, because they forage for benthic invertebrates which do not show a daily rhythm in abundance. Preliminary observations in three ducks found that the day-night differences in dive durations were retained after several days in constant darkness (36).

In conclusion, this study has shown that canvasback ducks exhibit significant day-night differences in their respiratory responses to progressive asphyxia, which are correlated with circadian rhythms in diving behavior. Ducks dive most often, and for longer, at night when respiratory sensitivity to combined hypoxia and hypercapnia is lower. Further work is required to determine whether these observations bear any direct causal relationship and, if so, which mechanisms mediate these effects. The data suggest that diving performance may be facilitated by reduced respiratory sensitivity to progressive asphyxia and, furthermore, that the circadian timing system may serve to synchronize the behavioral and respiratory control systems in birds.

Perspectives

By providing a temporal organization to both behavior and physiology, the circadian timing system may serve to optimize overall locomotor performance and thereby enhance the evolutionary fitness of semiaquatic diving species.