Altered pulmonary epithelial permeability in canine radiation lung injury

A radioaerosol scanning technique measuring regional clearance of sodium pertechnetate (99mTcO-4) and 99mTc-labeled diethylenetriaminepentaacetate (99mTc-DTPA) was used to assess changes in canine pulmonary epithelial permeability following lung irradiation. Doses of 2,000 cGy (11 dogs), 1,000 cGy (2 dogs), and 500 cGy (2 dogs) were given in one fraction to either the entire right hemithorax (500 cGy) or the right lower lung (1,000 and 2,000 cGy). Radioaerosol scans, chest roentgenograms, and computerized tomograms (CT) were obtained before and serially after irradiation. A dose of 2,000 cGy resulted in a decrease in regional pulmonary epithelial permeability to both 99mTcO4- and 99mTc-DTPA; both showed significant decreases from the 2nd wk postirradiation onward. In comparison, CT and chest roentgenogram did not become abnormal until 7.1 +/- 2.8 (SD) and 8.2 +/- 2.6 wk, respectively. Doses of 1,000 and 500 cGy produced reversible decreases in 99mTcO4- clearance. Lung morphology showed definite changes of radiation pneumonitis after 2,000 and 1,000 cGy but not after 500 cGy at approximately 9, 17, and 12 wk postirradiation, respectively. These results suggest that dose-dependent changes in pulmonary physiology may precede obvious structural alterations in radiation lung injury.     

Epithelial and endothelial flux after bypass in dogs: effect of positive end-expiratory pressure

Cardiopulmonary bypass (CPB) causes lung injury that occasionally progresses to the adult respiratory distress syndrome (ARDS). We measured the effect of 10 cmH2O of positive end-expiratory pressure (PEEP) on small solute and protein flux in dogs 1 wk before and 2 h after the completion of CPB. As an index of alveolar epithelial permeability, the clearance from lung to blood of inhaled technetium-99m-labeled diethylenetriaminepentaacetic acid (99mTc-DTPA) was measured. To assess microvascular endothelial integrity, the rate of accumulation in the lung interstitium of intravascular 113mIn-transferrin was measured. The clearance half time (t 1/2) for 99mTc-DTPA in the study dogs declined from 18.8 +/- 1.9 min (mean +/- SE) at base line to 9.4 +/- 2.0 min during PEEP (P less than 0.05). Two hours after CPB, the t 1/2 was 8.1 +/- 1.6 min at base line and unchanged during PEEP. The 113mIn-transferrin rate of accumulation was unchanged by PEEP before CPB. After CPB, the index was 3.25 +/- 0.95 slope/min X 10(-3) (P less than 0.05). Of the five dogs with a significant slope, four showed a decrease in microvascular flux during PEEP, although for the group the mean change in slope was not significant (P = 0.10). We conclude that the application of PEEP does not increase 99mTc-DTPA clearance in lungs already injured by CPB, and may actually decrease the apparent microvascular protein flux in some cases.     

Relationship of end-expiratory pressure, lung volume, and 99mTc-DTPA clearance

We investigated the dose-response effect of positive end-expiratory pressure (PEEP) and increased lung volume on the pulmonary clearance rate of aerosolized technetium-99m-labeled diethylenetriaminepentaacetic acid (99mTc-DTPA). Clearance of lung radioactivity was expressed as percent decrease per minute. Base-line clearance was measured while anesthetized sheep (n = 20) were ventilated with 0 cmH2O end-expiratory pressure. Clearance was remeasured during ventilation at 2.5, 5, 10, 15, or 20 cmH2O PEEP. Further studies showed stepwise increases in functional residual capacity (FRC) (P less than 0.05) measured at 0, 2.5, 5, 10, 15, and 20 cmH2O PEEP. At 2.5 cmH2O PEEP, the clearance rate was not different from that at base line (P less than 0.05), although FRC was increased from base line. Clearance rate increased progressively with increasing PEEP at 5, 10, and 15 cmH2O (P less than 0.05). Between 15 and 20 cmH2O PEEP, clearance rate was again unchanged, despite an increase in FRC. The pulmonary clearance of aerosolized 99mTc-DTPA shows a sigmoidal response to increasing FRC and PEEP, having both threshold and maximal effects. This relationship is most consistent with the hypothesis that alveolar epithelial permeability is increased by lung inflation.     

Reflex changes in tracheal smooth muscle tone during high-frequency oscillation

We examined the effect of high-frequency oscillatory ventilation (HFOV) on tracheal smooth muscle tension and upper airway resistance in anesthetized dogs. The animals were ventilated via a low tracheostomy by HFOV or conventional intermittent positive pressure ventilation (IPPV) with and without added positive end-expiratory pressure (PEEP). The transverse muscle tension of the trachea above the tracheostomy was measured and found to be lower during HFOV when compared with IPPV or IPPV with PEEP. When both vagi were cooled to 8 degrees C to interrupt afferent traffic from the lungs, there was no longer any difference between the modes of ventilation. In a second series of experiments, the airflow resistance of the upper airway above the tracheostomy was measured (Ruaw). During HFOV, Ruaw was significantly lower than during either IPPV or IPPV with PEEP. We conclude that HFOV induces a relaxation of tracheal smooth muscle and a reduction of upper airway resistance through a vagally mediated mechanism.     

Atrial reflexes during high frequency oscillatory ventilation

High frequency oscillatory ventilation (HFOV) is a new method of artificial ventilation which has been advocated for use in critically ill individuals. It alters the discharge in pulmonary stretch receptors (SAR) from a phasic to a continuous pattern. Since some cardiovascular neurones in the medulla are influenced by the discharge from SAR, experiments were undertaken to determine whether the reflexes from the left atrial (volume) receptors (LAR) were influenced by HFOV. The reflex increases in heart rate and urine flow which result from activation of the (LAR) were examined during both intermittent positive pressure ventilation (IPPV) and HFOV. In five dogs, the increase in heart rate was 23.9 +/- 4.3 and 24.5 +/- 5.4 beats/min during IPPV and HFOV, respectively. In six dogs the response of an increase in urine flow was examined and this response also was not altered by HFOV. It is concluded that the integrity of these reflexes was unaffected by HFOV in the anesthetized dog model.     

Effects of sustained exercise on pulmonary clearance of aerosolized 99mTc-DTPA

The effects of intensive prolonged exercise on the pulmonary clearance rate of aerosolized 99mTc-labeled diethylenetriaminepentaacetate (99mTc-DTPA) and pulmonary mechanics were studied in seven healthy nonsmoking volunteers. 99mTc-DTPA clearance and pulmonary mechanics (lung volumes and compliance) were assessed before and after 75 min of constant-load exercise performed on a treadmill, corresponding to 75% of maximal O2 uptake. Because both clearance measurements were made in similar conditions of pulmonary blood flow, respiratory rate, and tidal volume, changes in clearance rate can be assumed to represent changes of alveolar epithelial permeability. After exercise, total, apical, and basal clearance were significantly increased (P less than 0.01, 0.05, and 0.05, respectively) and the increases in total clearance and tidal volume observed during exercise were significantly correlated (P less than 0.05). In contrast, no significant change was found in pulmonary mechanics. These results show that prolonged intensive exercise induces an increase in epithelial permeability, which appears to be related to the mechanical effects of sustained increased ventilation. Because no change was evidenced in pulmonary volumes or in lung elasticity, our results suggest that this increase may result from alteration of the intercellular tight junctions rather than from a surfactant deficiency.     

Effects of high-frequency oscillatory ventilation on vagal and phrenic nerve activities

This study was undertaken to define the mechanism for the respiratory inhibition observed during high-frequency oscillatory ventilation (HFOV). The effects of HFOV on the activities of single units in the vagus (Vna) and phrenic nerves (Pna) were examined in pentobarbital-anesthetized dogs. The animals were either ventilated by intermittent positive-pressure ventilation (IPPV) with and without positive end-expiratory pressure (PEEP), or by HFOV at a frequency of 25 Hz and pump displacement volume of 3 ml/kg. In 13 vagal units the Vna was much higher during HFOV than during IPPV or airway occlusion at a matched airway pressure. Ten units in the phrenic nerves were examined, and Pna (expressed as bursts/min) was attenuated by HFOV in all of them. In four of them, the effect of cooling the vagi to 8-10 degrees C on Pna was examined, and it was found that HFOV failed to alter the Pna. We conclude that 1) HFOV stimulates the pulmonary vagal afferent fibers continuously and to a degree greater than that due to static lung inflation and increased airway pressure and 2) the increased vagal activity during HFOV probably causes phrenic nerve activity inhibition.     

Effect of inspiratory resistance and PEEP on 99mTc-DTPA clearance

Experiments were performed to determine the effect of markedly negative pleural pressure (Ppl) or positive end-expiratory pressure (PEEP) on the pulmonary clearance (k) of technetium-99m-labeled diethylenetriaminepentaacetic acid (99mTc-DTPA). A submicronic aerosol containing 99mTc-DTPA was insufflated into the lungs of anesthetized intubated sheep. In six experiments k was 0.44 +/- 0.46% (SD)/min during the initial 30 min and was unchanged during the subsequent 30-min interval [k = 0.21 +/- 12%/min] when there was markedly increased inspiratory resistance. A 3-mm-diam orifice in the inspiratory tubing created the resistance. It resulted on average in a 13-cmH2O decrease in inspiratory Ppl. In eight additional experiments sheep were exposed to 2, 10, and 15 cmH2O PEEP (20 min at each level). During 2 cmH2O PEEP k = 0.47 +/- 0.15%/min, and clearance increased slightly at 10 cmH2O PEEP [0.76 +/- 0.28%/min, P less than 0.01]. When PEEP was increased to 15 cmH2O a marked increase in clearance occurred [k = 1.95 +/- 1.08%/min, P less than 0.001]. The experiments demonstrate that markedly negative inspiratory pressures do not accelerate the clearance of 99mTc-DTPA from normal lungs. The effect of PEEP on k is nonlinear, with large effects being seen only with very large increases in PEEP.     

Effect of increased surface tension and assisted ventilation on 99mTc-DTPA clearance

Experiments were performed to determine the effects of conventional mechanical ventilation (CMV) and high-frequency oscillation (HFO) on the clearance of technetium-99m-labeled diethylenetriamine pentaacetate (99mTc-DTPA) from lungs with altered surface tension properties. A submicronic aerosol of 99mTc-DTPA was insufflated into the lungs of anesthetized, tracheotomized rabbits before and 1 h after the administration of the aerosolized detergent dioctyl sodium sulfosuccinate (OT). Rabbits were ventilated by one of four methods: 1) spontaneous breathing; 2) CMV at 12 cmH2O mean airway pressure (MAP); 3) HFO at 12 cmH2O MAP; 4) HFO at 16 cmH2O MAP. Administration of OT resulted in decreased arterial PO2 (PaO2), increased lung wet-to-dry weight ratios, and abnormal lung pressure-volume relationships, compatible with increased surface tension. 99mTc-DTPA clearance was accelerated after OT in all groups. The post-OT rate of clearance (k) was significantly faster (P less than 0.05) in the CMV at 12 cmH2O MAP [k = 7.57 +/- 0.71%/min (SE)] and HFO at 16 cmH2O MAP (k = 6.92 +/- 0.61%/min) groups than in the spontaneously breathing (k = 4.32 +/- 0.55%/min) and HFO at 12 cmH2O MAP (4.68 +/- 0.63%/min) groups. The clearance curves were biexponential in the former two groups. We conclude that pulmonary clearance of 99mTc-DTPA is accelerated in high surface tension pulmonary edema, and this effect is enhanced by both conventional ventilation and HFO at high mean airway pressure.     

Effects of positive end-expiratory pressure on hepatic blood flow and performance

Positive end-expiratory pressure (PEEP) may impair extrapulmonary organ function. However, the effects of PEEP on the liver are unclear. We tested the hypothesis that at a constant cardiac output (CO), PEEP does not induce changes in hepatic blood flow (QL) and parenchymal performance. In splenectomized, close-chested canine preparations (group I, n = 6), QL was derived as hepatic outflow using electromagnetic flow probes (QLemf), and hepatic performance was defined by extraction and clearance of indocyanine green (ICG). In a noninvasive model (group II, n = 7), the effects of PEEP on hepatic performance alone were similarly analyzed. Measurements were taken during intermittent positive-pressure ventilation (IPPV1), after addition of 10 cmH2O PEEP to IPPV (PEEP1), during continued PEEP but after return of CO to IPPV1 levels by intravascular volume infusions (PEEP2), and after removal of both PEEP and excess blood volume (IPPV2). Phasic inspiratory decreases in QLemf present during positive-pressure ventilation were not increased during either PEEP1 or PEEP2. Mean QLemf decreased proportionately with CO during PEEP1 (P less than 0.05), but was restored to IPPV1 levels in a parallel fashion with CO during PEEP2. The ICG pharmacokinetic responses to PEEP were complex, with differential effects on extraction and clearance. Despite this, hepatic performance was not imparied in either group. we conclude that global QL reductions during PEEP are proportional to PEEP-induced decreases in CO and are preventable by returning CO to pre-PEEP levels by intravascular volume infusions. However, covarying changes in blood volume and hepatic outflow resistance may independently modulate hepatic function.     

Lung epithelial permeability to aerosolized solutes: relation to position

The lung epithelial permeability to inhaled solutes is primarily attributed to the degree of distension of the interepithelial junctions and thus of the alveolar volume. To assess this hypothesis, a submicronic aerosol of technetium-99m-labeled diethylenetriamine pentaacetate (99mTc-DTPA) was inhaled by eight normal subjects in left lateral decubitus (LLD). The regional lung clearance of 99mTc-DTPA was measured in LLD, then in right lateral decubitus (RLD) to reverse the relative distension of the alveoli. Although in LLD the deposition of the aerosol is the greatest in the gravity-dependent regions of the left lung, their 99mTc-DTPA clearances are significantly lower than those of the nondependent regions of the right lung (0.7 +/- 0.3 vs. 2 +/- 0.8%/min, P less than 0.001). In RLD, these regions placed in opposite positions significantly reversed their clearances (1.6 +/- 0.8 vs. 0.6 +/- 0.2%/min, P less than 0.001). Results indicate in lateral decubitus a gravity gradient of 99mTc-DTPA clearances independent of the aerosol deposition. This gradient of epithelial permeability to solutes appears to be influenced by the gradient of alveolar volume.     

Effect of negative-pressure ventilation on lung water in permeability pulmonary edema

We have previously shown (Am. Rev. Respir. Dis. 136: 886-891, 1987) improved cardiac output in dogs with pulmonary edema ventilated with external continuous negative chest pressure ventilation (CNPV) using negative end-expiratory pressure (NEEP), compared with continuous positive-pressure ventilation (CPPV) using equivalent positive end-expiratory pressure (PEEP). The present study examined the effect on lung water of CNPV compared with CPPV to determine whether the increased venous return created by NEEP worsened pulmonary edema in dogs with acute lung injury. Oleic acid (0.06 ml/kg) was administered to 27 anesthetized dogs. Supine animals were then divided into three groups and ventilated for 6 h. The first group (n = 10) was treated with intermittent positive-pressure ventilation (IPPV) alone; the second (n = 9) received CNPV with 10 cmH2O NEEP; the third (n = 8) received CPPV with 10 cmH2O PEEP. CNPV and CPPV produced similar improvements in oxygenation over IPPV. However, cardiac output was significantly depressed by CPPV, but not by CNPV, when compared with IPPV. Although there were no differences in extravascular lung water (Qwl/dQl) between CNPV and CPPV, both significantly increased Qwl/dQl compared with IPPV (7.81 +/- 0.21 and 7.87 +/- 0.31 vs. 6.71 +/- 0.25, respectively, P less than 0.01 in both instances). CNPV and CPPV, but not IPPV, enhanced lung water accumulation in the perihilar areas where interstitial pressures may be most negative at higher lung volumes.     

Pulmonary clearance of radiotracers after positive end-expiratory pressure or acute lung injury

In anesthetized rabbits we measured clearance from lung to blood of eight aerosolized technetium-99m-labeled compounds: diethylenetriaminepentaacetate (99mTc-DTPA); cytochrome c; myoglobin; a myoglobin polymer; albumin; and anionic, cationic, and neutral dextrans of equivalent molecular size. We investigated the effect of applying positive end-expiratory pressure (PEEP) and, on a subsequent occasion, of injecting oleic acid intravenously to produce acute lung injury on the pulmonary clearance rate. Base-line clearance rates were monoexponential and varied with the molecular weights of the radiotracers. For each tracer the rate of clearance was increased a similar degree by either PEEP or oleic acid. However, with PEEP, clearance remained monoexponential, whereas after oleic acid, smaller molecular-weight radiotracers had multiexponential clearance curves. This suggests that after oleic acid the alveolar epithelium breaks down in a nonuniform fashion. We conclude that differentiation of the effect of PEEP from that of severe lung injury caused by oleic acid is not readily accomplished by either increasing the size of the tracer molecule or by varying the molecular charge.     

Effect of raised thoracic pressure and volume on 99mTc-DTPA clearance in humans

Although positive airway pressure is often used to treat acute pulmonary edema, the effects on epithelial solute flux are not well known. We measured independently the effect of 1) positive pressure and 2) voluntary hyperinflation on the clearance of inhaled technetium-99m-labeled diethylenetriaminepentaacetic acid (99mTc-DTPA) in six nonsmokers and six smokers. Lung volumes were monitored by inductance plethysmography. Each subject was studied in four situations: 1) low end-expiratory volume (LO-), 2) low volume plus 9 cmH2O continuous positive airway pressure (LO+), 3) high end-expiratory volume (HI-), and 4) high volume plus continuous positive airway pressure (HI+). The clearance half time of 99mTc-DTPA for the nonsmokers decreased from 64.8 +/- 7.0 min (mean +/- SE) at LO- to 23.2 +/- 5.3 min at HI- (P less than 0.05). Positive pressure had no synergistic effect. The mean clearance half time for the smokers was faster than nonsmokers at base line but unaffected by similar changes in thoracic volume and pressure. We conclude that, in nonsmokers, positive airway pressure increases 99mTc-DTPA clearance primarily through an increase in lung volume and that smokers are immune to these effects.     

Clearance of 99mTc-DTPA and experimentally increased alveolar surfactant content.

We measured clearance of 99mTc-labeled diethylenetriamine pentaacetic acid (99mTc-DTPA) in rabbits with experimentally increased alveolar surfactant content. In one group of animals, surfactant production was increased by treatment with ambroxol, and another group of animals was treated with tracheal instillation of natural surfactant. A group of untreated control animals and animals treated with instillation of saline were also studied. Clearance was measured during standard conditions of mechanical ventilation and during ventilation with large tidal volumes. In ambroxol- and surfactant-treated groups, clearance rate was reduced compared with untreated control animals. In contrast, clearance rate increased after saline instillation. The differences were observed at both modes of ventilation. The findings indicate that the pulmonary surfactant system is a rate-limiting factor for the clearance of 99mTc-DTPA and that the volume dependence of clearance is not explained by stretching of the alveolar wall only.     

Comparison of lung protection strategies using conventional and high-frequency oscillatory ventilation.

This study compared pathophysiological and biochemical indexes of acute lung injury in a saline-lavaged rabbit model with different ventilatory strategies: a control group consisting of moderate tidal volume (V(T)) (10-12 ml/kg) and low positive end-expiratory pressure (PEEP) (4-5 cmH(2)O); and three protective groups: 1) low V(T) (5-6 ml/kg) high PEEP, 2-3 cmH(2)O greater than the lower inflection point; 2) low V(T) (5-6 ml/kg), high PEEP (8-10 cmH(2)O); and 3) high-frequency oscillatory ventilation (HFOV). The strategy using PEEP > inflection point resulted in hypotension and barotrauma. HFOV attenuated the decrease in pulmonary compliance, the lung inflammation assessed by polymorphonuclear leukocyte infiltration and tumor necrosis factor-alpha concentration in the alveolar space, and pathological changes of the small airways and alveoli. Conventional mechanical ventilation using lung protection strategies (low V(T) high PEEP) only attenuated the decrease in oxygenation and pulmonary compliance. Therefore, HFOV may be a preferable option as a lung protection strategy.     

Pulmonary clearance of three aerosolized solutes in oleic acid-induced lung injury

We studied the effects of oleic acid (OA) on pulmonary clearance of three aerosolized radioactive solutes: 99mTc-diethylenetriamine pentaacetate (99mTc-DTPA), 67Ga-desferoxamine (67Ga-DFOM), and 111In-transferrin (111In-TF). Either 0.09 ml/kg OA or an equivalent volume of 0.9% NaCl (controls) was administered intravenously to 48 anesthetized, paralyzed dogs. Each animal received one aerosolized solute either 60 min after (protocol A) or 30 min before (protocol B) the infusion of OA or NaCl. In protocol A clearances of all three solutes were similar in OA and control animals. In contrast, in protocol B clearances of all three solutes increased significantly during OA infusion; during the next 60 min clearances of 99mTc-DTPA and 67Ga-DFOM returned to control values but 111In-TF remained increased. We conclude that 1) in OA-induced permeability edema pulmonary clearance of aerosolized solutes is increased when the aerosol is delivered 30 min before but not 60 min after injury, and 2) increased clearance persists only for large molecules, presumably because smaller molecules cross injured epithelium quickly and completely. These phenomena are best explained by a nonhomogeneous distribution of OA-induced injury.     

Collateral ventilation during high-frequency oscillation in dogs

Mechanics of collateral channels during high-frequency oscillatory ventilation (HFOV) were assessed in eight anesthetized dogs, using a modification of Hilpert's technique. Base-line functional residual capacity was measured with a body plethysmograph, with inspiratory efforts induced by phrenic nerve stimulation. The resistance (Rcoll) and time constant (Tcoll) of collateral channels at five lung volumes were measured during HFOV and positive end-expiratory pressure (PEEP). Rcoll and Tcoll were significantly higher during HFOV (P less than 0.001); the differences did not correlate with resting lung volumes. The calculated static compliance of the wedged segment was similar during HFOV and PEEP (P greater than 0.005). Mean pressures measured in small airways during HFOV corresponded to the midline between the inflation and deflation limbs of the static pressure-volume curves, indicating similar pressure-volume characteristics of the respiratory system during HFOV and static conditions. We conclude that HFOV increases resistance to gas flow through collateral channels but that this pathway may still be important in gas exchange.     

Comparison of three tracers for detecting lung epithelial injury in anesthetized sheep

We compared the ability of three aerosolized tracers to discriminate among control, lung inflation with a positive end expired pressure of 10 cmH2O, lung vascular hypertension and edema without lung injury, and lung edema with lung injury due to intravenous oleic acid. The tracers were 99mTc-diethylenetriaminepentaacetate (99mTc-DTPA, mol wt 492), 99mTc-human serum albumin (99mTc-ALB, mol wt 69,000), and 99mTc-aggregated albumin (99mTc-AGG ALB, mol wt 383,000). 99mTc-DTPA clearance measurements were not able to discriminate lung injury from lung inflation. The 99mTc-AGG ALB clearance rate was unchanged by lung inflation and increased slightly with lung injury. The 99mTc-ALB clearance rate (0.06 +/- 0.02%/min) was unchanged by lung inflation (0.09 +/- 0.02%/min, P greater than 0.05) or 4 h of hypertension without injury (0.09 +/- 0.04%/min, P greater than 0.05). Deposition of 99mTc-ALB within 15 min of the administration of the oleic acid increased the clearance rate to 0.19 +/- 0.06%/min, which correlated well with the postmortem lung water volume (r = 0.92, P less than 0.01). This did not occur when there was a 60-min delay in the deposition of 99mTc-ALB. We conclude that 99mTc-ALB is the best indicator for studying the effects of lung epithelial injury on protein and fluid transport into and out of the air spaces of the lungs in a minimally invasive manner.     

Effects of continuous positive-pressure ventilation in experimental pulmonary edema.

We compared the effects of continuous positive-pressure ventilation (CPPV), using 10 cmH2O positive end-expiratory pressure (PEEP), with intermittent positive-pressure ventilation (IPPV), on pulmonary extravascular water volume (PEWV) and lung function in dogs with pulmonary edema caused by elevated left atrial pressure and decreased colloid osmotic pressure. The PEWV was measured by gravimetric and double-isotope indicator dilution methods. Animals with high (22-33 mmHg), moderately elevated (12-20 mmHg), and normal (3-11 mmHg) left atrial pressures (Pla) were studied. The PEWV by both methods was significantly increased in the high and moderate Pla groups, the former greater than the latter (P less than 0.05). There was no difference in the PEWV between animals receiving CPPV and those receiving IPPV in both the high and moderately elevated Pla groups. However, in animals with high Pla, the Pao2 was significantly better maintained and the inflation pressure required to deliver a tidal volume of 12 ml/kg was significantly less with the use of CPPV than with IPPV. We conclude that in pulmonary edema associated with high Pla, PEEP does not reduce PEWV but does improve pulmonary function.