Exercise in Merino Sheep- the Relationships Between Work Intensity ...

Aust. J. Agric. Res., 1991, 4 2 , 599-620

Exercise in Merino Sheep- the Relationships Between Work Intensity, Endurance, Anaerobic Threshold and Glucose Metabolism D. W Pethick, C. B. Miller and N . G. Harman Division of Veterinary Biology, School of Veterinary Studies, Murdoch University, Murdoch, Western Australia, 61 50.

Abstract

The effect of exercise intensity on (i) the ability of sheep to sustain exercise and (ii) glucose metabolism was investigated in fed non-pregnant adult Merino ewes. Five animals were prepared with cannulae to study the splanchnic tissues using the arteriovenous difference technique either at rest or during 8 levels of exercise: 3, 5, 7 and 9 km h-I at either 0" or 9' incline. The anaerobic threshold, determined by elevation of blood lactate concentration or lactate/pyruvate ratio, occurred at a work rate of about 6-10 watts/kg body wt (7 km h-' on 0' incline, 3 km h-' on 9" incline). Only exercise well in excess of the anaerobic threshold resulted in ewes showing fatigue. Fatigue was not associated with carbohydrate depletion or lacticacidosis. Changes in the partial pressure of CO;! and the pH of blood indicated a marked respiratory alkalosis that was related to the severity of exercise, suggesting that thermoregulation may have been an important component of fatigue. Splanchnic blood flow declined when the intensity of exercise exceeded the anaerobic threshold; however, this did not compromise splanchnic function as assessed by oxygen and metabolite uptake. During exercise below the anaerobic threshold euglycemia was maintained while a pronounced hyperglycemia, that became more severe as the work rate increased, was found for exercise above the anaerobic threshold. The release of glucose by the liver increased significantly at all work rates and markedly so after the anaerobic threshold, such that the resultant hyperglycemia was consistent with an exaggerated hepatic glucose release due to 'feed forward' control. The contribution of lactate and glycerol to gluconeogenesis, assuming complete conversion, remained constant at 18-25% except at the highest work load where the contribution significantly declined to 9%. The decline was due to (i) saturation of hepatic lactate uptake and (ii) a failure for glycerol concentration and so uptake to increase beyond a work rate of 22 W kg-'. The requirement for gluconeogenic end products of digestion for animals grazed under extensive conditions would be 9-30% greater than for animals not exercising, depending upon the speed and inclination of exercise.

Introduction The ability of sheep to exercise over a wide range of work loads has not been studied. In particular the anaerobic threshold has not been determined. This threshold marks the level of exercise above which aerobic energy production must be supplemented by anaerobic metabolism, causing an increased lactate production (Wasserman 1986). The importance of this threshold for exercise performance is that (i) anaerobic carbohydrate catabolism can provide nearly twice as much power for muscle contraction when compared to aerobic utilization and (ii) glycogen utilization rapidly increases at this intensity, due

D. W. Pethick et al.

to the lowered efficiency of ATP production per mole of carbohydrate used; thus decreasing the potential duration of exercise (Sahlin 1986). Exercise has profound effects on the metabolism of glucose in sheep. Judson et al. (1976) found the entry rate of glucose to increase two-fold and the rate of gluconeogenesis to increase 1 . 5 times in animals walking at 5 km h-l, 0" incline. Increases in the entry rate of glucose have also been reported by Brockman (1979a) and Brockman and Halvorson (1981, 1982). The increased glucose availability allows for a 3-4-fold increase in glucose consumption by skeletal muscle in walking sheep (Bird et al. 1981; Pethick et al. 1987). These studies have highlighted the importance of glucose as a substrate during exercise. However, the effect of differing work rates on the metabolism of glucose has yet to be investigated in sheep. The aims of this study were to (1) determine the anaerobic threshold of sheep and (2) study the influence of work rate on (i) the ability of sheep to sustain exercise and (ii) the metabolism of glucose and gluconeogenic precursors.

Methods Animals and Diet Five mature non-pregnant, non-lactating Merino ewes with a 1-3 cm wool cover and weighing 34-40 kg ( 3 8 k 2 . 5 kg, mean2e.m.) were used. The diet was 700 g of pelleted ground lucerne (Medicago sativa) fed semi-continuosuly as 12 equal portions at intervals of 2 h. Feed analysis showed 14.2% protein and 29% crude fibre. Animal Preparation Catheters were implanted into the carotid artery and jugular, portal, hepatic and mesenteric veins essentially as described by Katz and Bergman ( 1 9 6 9 ~ ) .Splanchnic catheters were made from treated polyethylene ureteric catheters (Portex, Hythe, Kent, U.L.). Catalogue numbers 300/500/040, 300/500/050 and 300/500/060 were used to make the mesenteric, portal and hepatic vein catheters respectively. These catheters were used since they proved extremely durable and because they were radio-opaque. The catheters maintained their flexibility for many months, whereas catheters made from polyethylene (Dural Plastics, Dural, Australia) lost their flexibility and broke at the implantation site when sheep were exercised on a regular basis. Two catheters were placed into the portal vein to increase the likelihood of sampling blood over extended periods. One catheter was positioned such that the tip lay at the portal hiatus and the other was passed to the left branch of the portal vein so as the tip lay 7-8 cm from the portal hiatus. When necessary, clots were removed from the catheter tips using a teflon coated seldinger wire (Cook, Bloomington IN 47402, U.S.A.); catheters were not regularly flushed with heparin saline. Two mesenteric vein catheters were also implanted so as to increase the period for which at least one catheter remained in the mesenteric vein. The mesenteric catheter was the most likely to be excluded from the vein by tissue reaction thus making blood flow measurements by dye dilution invalid. Two weeks after surgery and weekly during experimentation, catheter location was checked radiologically utilizing an image intensifier. It was possible to check that radio-opaque dye injected into the mesenteric vein passed quickly over the portal venous catheter tips. Any obstruction to dye flow correlated with spurious blood flow data and so these animals were excluded from experimentation. An additional check of catheter location was performed at post mortem. Ewes were obtained from the Veterinary School farm and surgery was performed within 1 week. This short period of adjustment was chosen so that ewes would be comparable with animals in the field with respect to their state of training. One week after recovery they were placed in metabolism cages which were located in the room where all experiments

Exercise in Merino Sheep

occurred. A further one week was allowed for recovery followed by a 2 week period of training on a belt treadmill (3.6 mxO. 8 m; Scale and Engineering, Perth, Austarlia) at 5 km h-' on 9" incline for 45 min. During training, regular handling assured that ewes became accustomed to the experimental procedure. Experimental Procedure Each of the five sheep were subjected to the same set of 9 experiments, consisting of (i) one experiment at rest in the metabolism cage; (ii) experiments at each of 4 speeds: 3, 5, 7 and 9 km h-' at both 0" and 9" incline. In week 1, exercise at 5,3, 7 and 9 km h-I on 0" incline occurred on days 1, 2, 3 and 5 respectively. In week 2, an experiment at rest was conducted on day 3 and during week 3, the proctocol was similar to week 1 except that exercise was on a 9" incline. The protocol was reversed for two of the sheep. This design was utilized to aid repletion of glycogen reserves following exercise at the higher speeds (7 km h-' and above) where some degree of depletion would be expected. The ability of ruminant tissues to replete glycogen reserves after exercise is not known but repletion in the muscle tissue of cattle can take from 2-7 days (McVeigh et al. 1982). A longer experimental period may have been desirable but maintenance of catheter patency prompted the relatively short recovery times. Ewes were prevented from lagging by hanging a broom handle from the ceiling so as it hung just to the rear of the sheep; as soon as the sheep felt the handle they would move forward. Sheep were judged exhausted when they no longer responded to the stimulus. In all experiments, sampling began at 9.00 am, 60 mins after the beginning of the para-aminohippuric acid infusion. At rest, 10 simultaneous blood samples were collected from the carotid artery and portal and hepatic vein with sampling every 30 mins. During exercise at speeds below 9 km h-l a similar set of blood samples were taken every 2 . 5 mins with the remaining samples taken every 5 mins. Exercise lasted for 45 mins. For exercise at 9 km h-' the first 8 blood samples were taken every 2 . 5 mins with subsquent samples every 5 mins. Exercise on 0" and 9" was terminated after 30 and 25 mins respectively. At each sampling time 3 mL of blood was collected into heparinized syringes and immediately placed on ice. Within 60 mins of collection, 2 mL of blood was added to 3 mL of 5% perchloric acid for the subsequent determination of metabolite concentration. A further 0 . 5 mL was added to 5 - 5 mL of 10% trichloroacetic acid for the analysis of para-aminohippuric acid. Blood volume was determined gravimetrically. The remaining blood was used for analysis of blood gas content and pH within 60 min of collection. Assays Blood flow in the portal and hepatic vein was determined by the dye dilution principle of Katz and Bergman (1969b). The precision of para-aminohippuric acid measurement was improved by (i) ensuring that very close to 0 . 5 mL of blood was placed into 5.5 mL of trichloroacetic acid and (ii) by performing the boiling step in test tubes with screw caps (Herdt et a/. 1988). The concentration of glucose in all samples was determined on the perchloric acid filtrates using a Boehringer kit (Cat. No. 124036, Boehringer Mannheim, Sydney, Australia). The remaining metabolite assays were performed on neutralized perchloric acid filtrates. Lactate and glycerol were measured in all samples; lactate by the method of No11 (1985) and glycerol by the method of Garland and Randle (1962). Pyruvate was measured on arterial samples by the method of Czok and Lamprecht (1974), care was taken to analyse pyruvate immediately after neutralization of the perchloric acid filtrates. All enzymes and nucleotides were purchased from Boehringer (Boehringer Mannheim Aust. Pty Ltd), other chemicals were obtained from Sigma (Sigma Chemical Co., St Louis, MO 63178, U.S.A.). The content of oxygen in all blood samples was calculated from the haemoglobin content and oxygen saturation of haemoglobin measured photometrically using an 0SM2 Hemoximeter (Radiometer, Copenhagen). The calculation of oxygen content did not allow for dissolved oxygen (i.e. that not associated with haemoglobin). The pH and partial pressure of carbon dioxide (Pcoz) along with the calculated values of bicarbonate concentration were determined using a Corning 168 pH/Blood Gas Analyser (Corning, Halstead, Essex. U.K.).


Calculations

The net flux of nutrients across the gastrointestinal tract and the liver was determined utilizing the principles of the arteriovenous difference technique (Bergman 1975). The following flux rates were calculated. (a) If G = the net utilization of a metabolite by the gut (m mol h-l) then

where [A] and [PI are the concentrations (m mol L-l) of the metabolite in carotid arterial and portal vein blood respectively and Q p is the rate of blood flow (L h-l) in the portal vein. (b) If H =the net utilization of a metabolite by the lvier (m mol h-l) then

where [HI is the concentration (m mol L-l) of the metabolite in the hepatic vein blood and QA and QH are the rates of blood flow (L h-l) in the hepatic artery and hepatic vein respectively. The energy cost of exercise can be calculated from the equation of Brockway and Boyne (1980). If C is the energy cost of exercise (J kg-' m-l) then

where I is the inclination

(")

and S is the speed (rnmin-I). If W is the work rate (W kg-') then

This relationship was derived from measuring oxygen consumption over the range of 2.7-4.5 km h-l on 0"-10" incline. The equation does not appear to be valid above these limits since as the speed of exercise is increased work rate does not reach a maximum. It would be expected that oyygen consumption (and the derived work rate) would reach a maximum (VOzmax) as the speed of exercise is increased. Consequently the results of this paper are compared to speed and inclination on most occasions.

Statistics

Significant differences between time of sampling were determined by repeated measures analysis of variance using the statistical package SPSSX (SPSS Inc., Chicago, U.S.A.) Plateau values for blood flow and the rates of metabolite utilization by the gastrointestinal tract and liver, when calculated, were either (i) the mean of all values at each sampling time beyond 5 min when no significant time dependent changes were found or (ii) the mean of those values which ceased to be significantly different. All other comparisons utilized a one-way analysis of variance and Duncan's multiple range test.

Results Qualitative Assessment of Work Intensity Exercise at 3 km h-I represented a slow walk; at 5 km h-l a brisk walk; at 7 km h-I a trot while at 9 km h-I sheep began to gallop. Animals showed no signs of exhaustion or distress when walking. However, for exercise at 7 and 9 km h-l, particularly at 9" incline, the sheep became tired and were panting vigorously.


Blood Gasses and pH

The concentration of bicarbonate showed a significant decline as the speed of exercise increased beyond 7 and 3 km h-l at 0" and 9" incline respectively (Fig. 1a). Little change in the concentration of bicarbonate was seen after 10 min of exercise. Exercise at above 3 km h-I resulted in a significant hypocapnia that became progressively greater a s the work load increased (Fig. l b ) . At these rates of exercise (i.e. above 3 km h-l), the decline in PCOZcontinued throughout the exercise period.

7.31 0

, ,

I

, , ,

,

I

3

5

7

9

,

Speed (krn h-1)

Fig. 1. The effect of exercise intensity on blood gas parameters and pH of arterial blood. (a) Bicarbonate concentration, (b) partial pressure of carbon and 9" incline, W. dioxide and ( c ) pH. Exercise on 0" incline, -0 Values are meanks.e.m. of 5 sheep after 25 min exercise; *=significantly different to rest, P < 0.05.

A significant alkalemia was found at exercise above 3 km h-l except for the greatest work intensity (Fig. l c ) . The alkalemia became more severe during exercise at 7 km h-l, 0" incline and 5 km h-l, 9" incline and increased to pH 7 . 6 8 k 0 - 0 3 and pH 7 -61 rt0.02 respectively at the end of exercise.


Hepatic Blood Flow and Oxygen Consumption The total hepatic blood flow and contribution of the hepatic artery did not change during the period of exercise. The contribution of the hepatic artery to total hepatic blood flow was 14 5 k 2.0% and 1 3 . 3 r 1 9% during rest and at exercise of all intensities respectively. Walking at 3 or 5 km h-l on 0" incline resulted in no change to hepatic blood flow; however, all other intensities showed a significant reduction in blood flow (Fig. 2). Exercise on a 9" incline resulted in a significantly lower blood flow than at the corresponding speed at 0" incline at all speeds except 7 km h-I .

-

-

Speed (krn h-I)

Fig. 2. The effect of exercise intensity on total hepatic blood flow. Exercise on 0" icline, C---0 and 9" incline, 9--@. Values, shown as mean?s.e.m., represent the mean of all values at each sampling time beyond 5 min exercise for 5 sheep. There were no significant time dependant changes. *=significantly different to rest, P i 0.05.

There was a tendency for oxygen consumption by the splanchnic tissues to increase during exercise; however, this increase was not consistently significant (Table 1). The maintained or increased consumption of oxygen was due to a large increase in the extraction of oxygen. For example, during exercise at 9 km h-l on 9" incline, both tissues increased the arteriovenous difference for oxygen by 2 . 2 times. Table 1. The Effect of Exercise Intensity on Splanchnic Oxygen Consumption (m mol h-1) Values, shown as meancs.e.m., represent the mean of all values at each sampling time beyond 5 min of exercise for 5 sheep. There were no significant time dependant changes. The inclination of exercise was 0" or 9"

Speed (km h-'1 Rest 3 5 7 9 A

Liver

Gut 0" 79 2 7 69 2 7 8423 81 +-6 95 2 7

9"

6529 7622 104214 110216~

Significantly different to rest, P < 0.05 values.

0" 7226 75 + 4 100 t 5 10727 116210~

9"

8029 10625 143216~ 108221


Anaerobic Threshold Arterial lactate concentration

The influence of time on the concentration of lactate in arterial blood during exercise is shown in Fig. 3. An immediate increase is evident at all work intensities; however, this was only sustained at the higher work loads where the concentration tended to either stabilize or increase slowly with time. As the speed of exercise increased, lactate concentration rose in a curvilinear manner, with significant accumulation at 7 and 5 km h-l for 0" and 9" incline respectively (Fig. 4). This corresponds to a work rate of 10 and 12 w kg-' respectively. The pattern of change in lactate concentration was related to the changes seen for bicarbonate concentration in blood. Thus both parameters did not show large time dependent changes after the early phase of exercise and the increase in lactate concentration was inversely related to the concentration of bicarbonate (Spearman's r= 0.98). Lactate/pyruvate ratio

At the onset of exercise the lactate/pyruvate ratio showed a sharp increase which returned to a lower plateau value after 10 min. The plateau values, plotted against the speed of exercise are shown in Fig. 5. The pattern of increase with respect to speed or work rate was similar to that of lactate concentration; however, a significant increase in the ratio was seen at a lower work rate of 6 . 5 W kg-l. Glucose Metabolism Blood glucose

The time dependant changes in glucose concentration at different levels of exercise are shown in Fig. 6. At all levels of exercise there was an increase in the concentration of glucose after a tendency for an initial decrease at 2 5 min. The increase was significantly only at speeds where the lactate concentration was significantly elevated.

-

Hepatic glucose release

Figure 7 shows the time dependant changes in the production of glucose by the liver at different speeds. At each exercise intensity, values rapidly reached a plateau. A significant increase in hepatic glucose production over the resting value was detected at each level of exercise. At each inclination, every speed was significantly different to the others and at each speed the effect of inclination was also significant. Variations were again most extreme at the higher intensities of exercise, thus at 9 km h-I the production of glucose increased 10 and 20 times at 0" and 9" incline respectively. Glucose utilization by the gut

The time course of the hyperglycemia seen during exercise was such that the concentration of glucose in blood had either reached or was approaching a plateau value (see Fig. 6). Therefore the extra glucose released by the liver


O" incline

r

Duration of exercise (mins)

Fig. 3. Changes in the concentration of lactate in arterial blood during exercise at different speeds on either 0" or 9" incline. Exercise at 3 km h-l, 0--0; 5 km h-l, W;7 km h-l, n-• and 9 km h-l, H.Values are meanks.e.m. for 5 sheep.

Speed (km h-I)

Fig. 4. The effect of exercise intensity on the concentration of lactate in arterial blood. Exercise on 0" incline, -0 and 9" incline, U. Values are mean+s.e.m. of 5 sheep after 25 min exercise; *=significantly different to rest, P < 0.05.


Speed (krn h-I) Fig. 5. The effect of exercise intensity on the molar ratio of lactate/pyruvate in arterial blood. Exercise o n 0" incline, 0-Oand 9" incline, W.Values, shown as mean+s.e.m., represent the mean of all values at each sampling time beyond 10 min exercise for 5 sheep. There were no significant time dependant changes after 10 min exercise. *=significantly different to rest, P < 0.05.

was being utilized by the extra hapatic tissues. Figure 8 provides evidence for this hypothesis since it shows a significant, linear relationship between blood glucose concentration and the utilisation of glucose by the gut. Gluconeogenic Precursors Lactate The utilization of lactate by the liver responded to circulating concentration up to 7 m mol L-l (Fig. 9a). This range of lactate concentration included all speeds except 9 km h-I at 9" incline. At this work load the utilization of lactate did not increase despite a further 2-fold increase in lactate concentration. In contrast, the utilization of lactate by the gut showed a linear relationship with the concentration of lactate in arterial blood (Fig. 9b). At concentrations below 2 m mol L-l there was a net production of lactate by the gut and above this there was a net utilization. The maximal potential contribution of lactate to gluconeogenesis assuming complete conversion of lactate to glucose is shown in Table 2. Lactate could account for 29% of the glucose production at rest. During exercise the mean contribution for all levels of work was 1 2 & 1 %with no significant differences between work loads. There was a tendency for decline at the highest work rate where lactate could account for 7% of the glucose released. Glycerol The ultilization of glycerol by the liver was linearly related to the arterial concentration with no maximum value being reached (Fig. 10). Uptake increased as a linear function of speed during work at 0" incline; however, during exercise at 9" incline the utilization increased up to 7 km h-l, after which there was a decline. The decline was due to a consistently lower concentration of glycerol in blood during exercise at 9 km h-l on 9" incline.

D.

W.Pethick et al.

O" incline 7.0 r

6.0

-

5.0

-

4.0

-

3.0

-

2.0

-

7

0

.-s' C 0

+

0 w

9" incline 0

9.0r

Duration of exorcise (rnins) Fig. 6. Changes in the concentration of glucose in arterial blood during exercise at different speeds on either 0" or 9" incline. Exercise at 3 km h-l, 0-0; 5 km h-l, .--a; 7 km h-l, 9 km h-l, U+. Values are mean2s.e.m. for 5 o-nand sheep. Table 2. The Effect of Exercise Intensity on the Potential Contribution of Lactate and Glycerol to Hepatic Gluconeogenesis (%) Values shown as mean1-s.e.m. represent the mean of all values at each sampling time beyond 5 min of exercise for 5 sheep. The calculation assumed that lactate and glycerol utilized by the liver was completely converted to glucose. The inclination of exercise was 0" or 9" Speed (km h-'1 Rest 3 5 7 9 A

Glycerol

Lactate

0"

9+ 1 13k3 1222 7+lA 61-lA

9"

0"

9"

1321 10+2 61-1 2+0.5*

2925 111-4 161-2 11+2 14 1- 2

111-1 171-1 11k1 71-1

-

-

Significantly different to 3 km h-I at the same inclination, P < 0.05.

The potential contribution of glycerol to gluconeogenesis, assuming complete coversion of glycerol to glucose is shown in Table 2. At rest glycerol could account for 9% of the glucose released. During exercise there was a trend for the role of glycerol as a gluconeogenic precursor to decline as the work


0' incline 140 [

I

0

5

10

15

20

25

30

35

40

45

50

Duration of exercise (rnins) Fig. 7. Changes in the net production of glucose by the liver during exercise at different speeds on either 0" or 9" incline. 5 km h-l, @--a; 7 krn h-', Exercise at 3 krn h-l, 0-0; o-• 9 km h-', H.Values are rnean2s.e.m. for 5 sheep.

load increased. At the lower work rates (3 km h-l) the contribution was equivalent to lactate at 13% but at the highest work intensity the contribution had declined to 2%, a value 3 . 5 times lower than for lactate. The potential contribution of both lactate and glycerol to gluconeogenesis remained relatively constant at an average 2 3 r 2 % for all except the highest work rate where it dropped significantly to 9%, due mainly to the decline in the uptake of glycerol by the liver.

Discussion Anaerobic Threshold The anaerobic threshold was estimated by two methods. Firstly by the determination of a significant increase in the concentration of lactate in arterial blood. This indicated the threshold to be at the work rate of 10-12 W kg-l. Changes in the bicarbonate content of blood, which largely reflect buffering of the protons from lactate (Wasserman et al. 1973), confirm these values. The second method was to determine at which work loads the lactate/pyruvate ratio increased. This method produced essentially similar results except that the threshold was found to be at a lower range of work rates at 6-5-10 W

D. W. Pethick et a/.

Arterial glucose concentration (mmol L") Fig. 8. The relationship between the concentration of glucose in arterial blood and the net utilization of glucose by the gut. El rest; O a n d a, exercise at 3 krn h-I on 0" and 9" incline respectively; nand., exercise at 5 krn h-l on 0' and 9" incline respectively; A and A, exercise at 7 krn h-I on 0" and 9" incline respectively; 0 and +, exercise at 9 krn h-I on 0" and 9" incline respectively. Values, shown as meancs.e.m., represent the mean of all values a t each sampling time beyond 5 rnin exercise for 5 sheep. If x is the concentration of glucose in arterial blood (rnrnol L-') and y the net ultilization of glucose by the gut (rnrnol h-I), then y=4.12x-8.33, r = 0 . 9 8 .

k g 1 . The anaerobic threshold of mature merino ewes is therefore at a work rate of about 10 W kgw1. The primary reason for lactate accumulation would seem to be a relative lack of oxygen since the lactate/pyruvate ratio increased before the concentration of lactate in blood (Wasserman 1986).

Exercise Capacity Fatigue is classified as either central or peripheral (Newsholme and Leech 1983). Central fatigue implies humeral factors which affect the motivation for exercise. Sensory information from contracting muscle fibres or thermoreceptors represent examples of information processed centrally which could induce fatigue. Peripheral fatigue implies a limitation in the motor unit. Typically it involves fuel (glycogen) depletion or end product (hydrogen ion) accumulation in the contracting muscle fibre. In this study sheep became fatigued only at speeds well in excess of the anaerobic threshold (9 km h-l, 0" incline and 7 and 9 km h-I 9" incline). Given the rapid use of glycogen which is known to occur at these work loads (Sahlin 1986), it is tempting to implicate glycogen depletion as the cause of fatigue.


Arterial lactate concentration (mmol L-') Fig. 9. The relationship between the concentration of lactate in arterial blood and the net rest; utilization of lactate. (a) Net utilization by the liver. (b) Net utilization by the gut. 0 and @, exercise at 3 km h-l on 0" and 9" incline respectively; and D, exercise at 5 km h-I on 0" and 9" incline respectively; A and A, exercise at 7 km h-I on 0" and 9" incline respectively; 0 and +, exercise at 9 km h-I on 0" and 9" incline respectively. Values, shown as mean+s.e.m., represent the mean of all values at each sampling time beyond 5 min exercise. If x is the concentration of lactate in arterial blood (mmol L-I) and y is the net utilization of lactate by the gut (mmol h-I), then y = 3 . 0 8 ~ - 6 . 9 9 , r = 0 . 9 9 . If y is the net utilization of lactate by the liver (mmol h-l) then y=13.1+21.1 (log x), r = 0 . 9 2 .


Arterial glycerol concentration (mmol L-') Fig. 10. The relationship between the concentration of glycerol in arterial blood and the net utilization of glycerol by the liver. El rest; Oand., exercise at 3 km h-I on 0' and 9" incline respectively; LY and W, exercise at 5 km h-I on 0" and 9" incline respectively; A and A, exercise at 7 km h-I on 0" and 9" incline respectively; 0 and +, exercise at 9 km h-' on 0" and 9" incline respectively. Values, shown a s mean+.s.e.m., represent the mean of all values at each sampling time beyond 5 min exercise for 5 sheep. If x is the concentration of glycerol in arterial blood (mmol L-I) and y the net ultilization of glycerol by the liver (mmol h-I), then y=50.38x-2 -01, r=O .93.

However, there is little evidence that carbohydrate depletion was associated with fatigue in this work. Hepatic glycogen reserves were not depleted since hypoglycemia was never observed and the rate of glucose production by the liver was constant throughout the period of exercise. Assuming the liver glycogen was initially 4 . 1 % (Pethick and Harman 1989), then 70-80% of the glycogen reserves would have been utilized to sustain the observed rates of hepatic glucose release (mean liver weight was 467 g). It is likely that less glycogen was utilized since gluconeogenesis could have accounted for 18-38% of the glucose release (see below). The reserves of glycogen in skeletal muscle should also be adequate since the glycogen content of liver becomes depleted before that of muscle in both sheep (Pethick and Harman 1989) and rats (Hickson et al. 1977). However, specific depletion of fast twitch muscle fibres cannot be ruled out in this study. End product accumulation in the form of hydrogen ions generated from the dissociation of lactic acid is an alternative hypothesis for exhaustion. This corresponds to the creatine phosphate threshold (Sahlin 1986) and at this point the concentration of lactic acid in contracting muscle continually increases and the levels of phosphocreatine and ATP decline. It is the classical explanation


for exhaustion seen during sprinting. The close relationship between the concentration of lactate and bicarbonate in the blood is evidence for a tendency toward systemic metabolic acidosis induced by the production of lactic acid above the anaerobic threshold; a response well documented in the human athlete (Wasserman 1986). However, the pattern of increase in blood lactate concentration (or decrease in bicarbonate concentration) was not suggestive of a fatigue inducing lactic acidosis at exercise of either 9 km h-', 0" incline or 7 km h-l, 9" incline. Lactic acid levels were similar after 30 min of exercise yet the sheep were able to exercise for a further 15 min at 7 km h-l on 9" incline. At 9 km h-l on 9" incline, the level of lactate in blood was substantial and more suggestive of a lactic acidosis in contracting muscle. However, the relatively stable concentration of lactate in blood suggests that production and release by muscle had reached an equilibrium and so was not increasing to finally reach a critical maximum. In addition the uptake of lactate by the liver declined in relation to arterial concentration at this work load indicating that in part the lactate accumulation was due to an inability to metabolize the extra lactate produced. Consequently, even at this work load lacticacidosis may not have been the cause of fatigue. Central fatigue initiated by hypertherirna is a further possibility. The pattern of decline in the Pco2 of blood is consistent with a respiratory induced hypocapnia that was more extreme as both the intensity and duration of exercise increased at work loads above 3 km h-l. Respiratory hypocapnia during exercise implies an increased alveolar ventilation over and above the drive for carbon dioxide excretion and is most likely due to a thermoregulatory drive. Hypoxia of arterial blood, an alternative cause of hyperventilation was not observed at any work rate. The drive for hyperventilation was sufficient to induce an alkalemia and so an overall respiratory alkalosis except at the highest work rate, where the pH was not different to rest, presumably due to the high lactic acid concentration. In contrast, exercise above the anaerobic threshold in the human athlete results in a tendency for systemic acidosis and no change in the P C O of ~ blood (Wasserman et al. 1973). Bell et al. (1983b) have shown that sheep with a 10 mm wool cover have difficulty in dissipating the thermal load of exercise such that at an ambient temperature of 40°C, 3 out of 5 animals became fatigued after 30 min of exercise at 2 . 6 km h-l on 10" incline. A significant role of panting as a mechanism for heat dissipation was emphasized by the authors. This was further supported by Pethick and Harman (1989) where exhaustion was associated with heavy panting and a pathological alkalemia in a sheep that could not sustain extended exercise at 4 . 5 km h-l on 9" incline. The results of this study extend the previous work and point to further studies of thermoregulation in sheep during exercise. Splanchnic Function

Exercise in sheep elicits an increase in the rate of blood flow to working muscle, adipose tissue, skin at the extremities and tissues of the upper respiratory tract. Adaptations to maintain this increased flow include increased cardiac output and vasoconstriction in several visceral organs (Bell et al. 1983b). In this study vasoconstriction in the splanchnic bed was detected at work rates at or above the anaerobic threshold. This reduction is probably

D. W.Pethick et a/.

mediated via the sympathetic (splanchnic) nerves of the gut (Barnes et al. 1986) which must become sufficiently effective only after the anaerobic threshold. The magnitude of the decline is similar to that found by Brockman (1987) at the corresponding work load. In general, the decline in blood flow would appear less dramatic than in the human athlete (Rowel1 1974) and may represent a significant adaptation by the sheep to maintain gut function; particularly to allow normal rates of nutrient absorption. Findings of Bell et al. (1983b) that blood flow to the rumen was maintained despite decreases in the flow of other gut compartments support this hypothesis. Liver and gut functions as judged by nutrient uptake were generally not compromised by the decline in blood flow seen during exercise. The liver and gut can increase the extraction of nutrients and so compensate for the decrease in blood flow. The tendency for a significant increase in oxygen consumption by the liver could be due to increased rates of gluconeogenesis (Judson et al. 1976) and urea synthesis. Exercise is known to increase the output of ammonia by working muscle (Katz et al. 1986). Increases in the consumption of oxygen by the gut indicate an increased metabolic rate that may have been associated with the disposal of extra glucose and lactate utilized during exercise. The utilization of glycerol by the liver and lactate and glucose by the gut was all first order with respect to substrate, again indicating normal splanchnic function during exercise. Lactate uptake by the liver did reach a maximum at the greatest work load and this value was considerably less than that found by Naylor et al. (1984) in resting sheep infused with lactate. Therefore, the uptake of lactate by the liver is partially compromised during exercise when compared to sheep at rest. Glucose Metabolism Feed forward control Hyperglycemia had been previously observed in sheep during exercise (Jarrett et al. 1976; Brockman 1979b; Bell et al. 1983a). However, this study extends this observatio$to show that euglycemia is maintained below the anaerobic threshold. Further, as the rates of exercise increase above this threshold the degree of hyperglycemia increases markedly. Increased release by the liver is the probable hypergly?emic agent since the uptake of glucose by the hindlimb continues to rise as @e rate of exercise increases past the anaerobic threhold (Jarrett et al. 1976). These responses suggest that at exercise below the anaerobk threshold there is normal 'feedback' regulation of hepatic glucose production such that increases in the production of glucose closely match the extra requirements of exercise. At higher work rates the production of glucose 4s greater than the rate of clearance. This phenomenon has been previously described during exercise in the rat and human where it has been called 'feedforward' control (Kjaer et al. 1987; Sonne et al. 1987; Winder e t al. 1988). Factors determining 'The degree of hyperglycemia seen upon exercise include level of feeding (Sonne et al. 1987) and work rate (Vissing et al. 1988). This study highlights the anaerobic threshold as the point at which 'feedforward' control is initiated. In addition it shows thatusheep represent a striking example of this regulation.


The mechanisms for 'feedforward' control of glucose metabolism are not well understood. However, it is supposed that activation of motor pathways in the central nervous system is accompanied by the activation of the sympathoadrenal response at the higher work rates. In sheep, both catecholamine (Palmer et 41. 1984) and glucagon (Brockman 1979b) response is related to the severity of exercise. The parallel response of these hormones is predictable since catecholamines stimulate glucagon secretion in sheep (Bassett 1972). In more comprehensive studies it has been shown that when human subjects proceed from low to high work rates, there is a disproportionately large increase in the catecholamine response (see Lehmann et al. 1981). Thus the hypothesis is that at levels of exercise above the anaerobic threshold, an exaggerated sympathoadrenal response results in an increased glucose output by the liver that is not readily influenced by the normal feedback signals. Clearly, estimates of circulating catecholamine in sheep subjected to different work rates are required for further substantiation. Carbohydrate utilization

The hyperglycemia seen upon intensive exercise may serve to stimulate the uptake of glucose by working muscle. Clearly tissue uptake of glucose was increasing to match the extra glucose released by the liver as the work rate increased. The tendency for blood glucose concentration to reach a plateau value despite a constant elevation in glucose production is evidence for this. Blood glucose was always below the renal theshold of 8-8-11 . l m mol L-l (McCandless et al. 1948) verifying that glucose was not simply being cleared into the urine. Certainly the gut responded to an increased glucose concentration by increased utilization despite a declining blood flow. This adaptation is potentially wasteful since it is a generalized response that would see increased glucose utilization by many tissues in addition to working muscle; however, it does offer a purpose for 'feedforward' control. A similar response was seen for lactate, such that tissue utilization increased to maintain a relatively constant lactate concentration in blood throughout the period of exercise. Several tissues are involved in the clearance of lactate. This study identifies the liver and the gut as important sites of lactate removal. The role of the gut as a net lactate utilizer becomes important only at relatively high concentrations of lactate, an adaptation which allows for continued lactate clearance despite saturation of hepatic uptake. Other tissues capable of lactate removal include skeletal muscle (Faichney et al. 1981) and the heart (Fisher et al. 1980). Hepatic glucose release

This study demonstrates that exercise is the most powerful physiological drive for hepatic glucose release yet to be reported. Rates (m mol h-l kg body weight-l) reported for pregnancy and lactation (Van der Walt et al. 1983) are below those found for exercise at 5 km h-l on 0" incline and the highest work rate of this study increased glucose production to 10 times that found for twin pregnant ewes. Interpretation of net glucose release warrants some caution due to a lack of steady state. Zeirler (1961) stressed the need for constant

D. W.Pethick et al.

rates of blood flow, tissue metabolism and circulating concentration when assessing nutrient exchange using the arteriovenous difference technique. Th rates of blood flow and tissue metabolism (as judged by oxygen uptake) were constant throughout the period of exercise. However, the arterial concentration of glucose did increase, especially at the higher work rates. For example, exercise at the highest work rate resulted in a 6% min-I change in the blood glucose concentration during the initial phase ( 7 . 5 - 1 2 . 5 min), but toward the end of exercise (20-25 min) the change had declined to 1%min-l. Despite these changes the estimated rate of glucose release by the liver rapidly reached a maximum and stayed constant throughout the period of exercise, an observation which supports the validity of the technique in this study. The basis of Zeirler's concern was the notion of a relatively long transit time for blood through tissues. Preliminary work by the authors estimating the time for radio-opaque dye to be cleared from the portal vein into the hepatic vein suggests that the transit time for blood is relatively short at about 7 seconds, an observation consistent with similar measurements in the resting forearm (Mottram e t al. 1973). Resting muscle might be expected to have a relatively long transit time for blood due to the relatively low blood flow when compared with the liver.

Gluconeogenesis The increased supply of the glucose during exercise must originate from either gluconeogenesis or glycogenolysis. The major gluconeogenic substrates are lactate, glycerol, propionate and glycogenic amino acids (Weekes 1979). Estimates of lactate and glycerol uptake show that gluconeogenesis from these substrates could potentially increase in proportion to work intensity except at the highest work rate. It is probable that propionate uptake by the liver and so its maximum potential contribution to gluconeogenesis would remain at the resting value of 1 7 -5 m mol h-l (Bergman and Wolff 1971). The results of Harman and Pethick (1987) can be used to approximate the gluconeogenic potential from amino acids both below and above the anaerobic threshold if it is assumed that alanine uptake by the liver represents 25% of the maximum potential contribution (Lindsay 1982). Using these assumptions and the data from this study, 145% of the glucose output at rest can be accounted for as gluconeogenic carbon utilized by the liver at rest, a value consistent with other work and highlighting that complete conversion of precursors to glucose is not necessary (Weekes 1979). When sheep are walked at either 0" or 9" incline the estimates range over 55-95%. Similar estimates in human athletes working at b e l o r the anaerobic threshold for 40 min show that gluconeogenesis could account for 23% of the glucose release (Ahlborg e t al. 1974). It is clear that gluconeogenesis is potentially of great significance in the walking sheep, an adaptation that would make them resistant to fatigue induced by carbohydrate depletion. The above calculations make assumptions that need further verification. Propionate absorption is unlikely to be markedly influenced by exercise, especially considering that splanchnic function was normal for all levels of exercise. Moreover, the entry rate of acetate does not change during exercise (Judson et al. 1976), again pointing to normal rates of absorption from the gut,


an interpretation which assumes no large changes in the rate of endogenous acetate production. Estimates for the contribution of amino acids are more open to error. Brockman (1987) found no increase in alanine uptake while Harman and Pethick (1987) found an increased uptake during exercise that was not significantly influenced by the intensity of exercise. The results of Brockway and Lobley (1982) would support those of Harman and Pethick (1987) since they found increased rates of protein catabolism in walking sheep. This work points to exercise as an important determinant of nutrient requirements. Merino sheep grazed under extensive management systems walk an average distance of about 7 km per day (range 5.6-17.8 km; Lynch 1967, 1974; Squires and Wilson 1971; Squires et al. 1972; Squires 1974, 1976). The speed of walking is poorly documented: however, if animals walk at 3 km h-l on a 0" incline then the daily glucose production by the liver would increase by 9% when compared to an inactive sheep, a value that does not take into account additional glucose production that may occur in the recovery period. If the same exercise was performed at 9" incline then glucose secretion would increase by 30%. The activity allowance set for sheep kept 'out-of-doors' is an extra 7% of metabolizable energy (table 3.31, Commonwealth Agricultural Bureax 1980). Clearly the allowance may not be enough for Merino sheep grazed under Australian conditions, especially if animals are consuming feed of low metabolizability. Fat Metabolism

The circulating concentration of glycerol is typically thought to represent the rate of lipolysis in adipose tissue. Evidence for this includes the low rate of glycerol kinase (EC 2.7.1.30) in adipose tissue (Shirley et al. 1973) which would allow quantitative release of glycerol and the linear relationship between glycerol concentration and entry rate (Dunshea et al. 1990). In contrast to lactate or glucose, the concentration of glycerol did not continue to increase at the higher work rates, suggesting that lipolysis had reached a maximum. This apparent limitation of fuel supply may not affect exercise performance because the maximum rate of long chain fatty acid oxidation probably occurs before the maximum rate of lipolysis. This is because lipolysis does not quantitatively reflect the rate of oxidation due to re-esterification of long chain fatty acids (Dunshea et al. 1990). In the human athlete, peak rates of oxidation occur at just below the anaerobic thresh01 (Sahlin 1986). In conclusion, this work has shown that sheep develop a respiratory alkalosis that is consistent with thermoregulation as a limiting factor for the exercise performance of sheep. The anaerobic threshold has been described and shown to be the point where glucose homeostasis switches from 'feedback' to 'feedforward' control. Exercise, even at low intensity, elicited an accelerated release of glucose by the liver that would require a significant increase in metabolizable energy intake to supply more gluconeogenic end products of digestion.

Acknowledgments The skilled technical assistance of Ms D. Wilson and Ms B. Olszewswki is gratefully acknowledged. Mr K. Chong is thanked for the preparation of


figures. Financial support for part of this work was provided by the Australian Research Council (ARC).

References Ahlborg, G., Felig, P., Hagenfeldt, L., Hendler, R., and Wahren, J. (1974). Substrate turnover during prolonged exercise in man. J. Clin. Invest. 53, 1080-90. Barnes, R. J., Comline, R. S., and Dobson, A. (1986). The control of splanchnic blood flow. In 'Control of Digestion and Metabolism in Ruminants'. Proc. 6th Inter. Syrnp. Ruminant Physiol., Banff, Canada, Sept. 1984. (Eds L. P. Milligan, W. L. Grovurn and A. Dobson.) pp. 41-59. (Prentice-Hall: Englewood Cliffs, New Jersey.) Bassett, J. M. (1972). Plasma glucagon concentrations in sheep: their regulation and relation to concentrations of insulin and growth hormone. Aust. J. Biol. Sci. 25, 1277-87. Bell, A. W., Bassett, J. M., Chandler, K. D., and Boston, R. C. (1983a). Fetal and maternal responses to exercise in the pregnant ewe. J. Develop. Physiol. 5, 129-41. Bell, A. W., Hales, J. R. S., King, R. B., and Fawcett, A. A. (1983b). Influence of heat stress on exercise-induced changes in regional blood flow in sheep. J. Appl. Physiol. 55, 1916-23. Bergman, E. N., and Wolff, J. E. (1971). Metabolism of volatile fatty acids by liver and portal-drovined viscera in sheep. Am. J. Physiol. 22, 586-92. Bergman, E. N. (1975). Production and utilisation of metabolites by the alimentary tract as measured in portal and hepatic blood. In 'Digestion and Metabolism in the Ruminant'. Proc. 4th Inter. Symp. Ruminant Physiol. (Eds I. W. McDonald and A. C. I. Warner.) pp. 292-305. (Uni. New England Publ. Unit: Armidale, N.S.W.) Bird, A. R., Chandler, K. D., and Bell, A. W. (1981). Effects of exercise and plane of nutrition on nutrient utilization by the hindlimb of the sheep. Aust. J. Biol. Sci. 34, 541-50. Brockman, R. P. (1979a). Effect of somatostatin on plasma glucagon and insulin, and glucose turnover in exercising sheep. J. Appl. Physiol. 47, 273-8. Brockman, R. P. (1979b). Glucagon responses to exercise in sheep. Aust. J. Biol. Sci. 32, 21 5-20. Brockman, R. P., and Halvorson, R. (1981). Effects of stomatostatin and adrenergic blocade on glucagon, insulin and glucose in exercising sheep. Aust. J. Biol. Sci. 34, 551-60. Brockman, R. P., and Halvorson, R. (1982). Glucose, glucagon and insulin during adrenergic blockade in exercising sheep. J. Appl. Physiol. 52(2), 315-9. Brockman, R. P. (1987). Effect of exercise on net hepatic uptake of lactate, pyruvate, alanine, and glycerol in sheep. Can. J. Physiol. Pharmacol. 65, 2065-70. Brockway, J. M., and Boyne, A. W. (1980). The energy cost for sheep of walking on gradients. In 'Energy Metabolism'. Proc. 8th Symposium on Energy Metablism, Cambridge. (Ed. L. E. Mount.) pp. 449-453. (Butterworths: London.) Brockway, J. M., and Lobley, G. E. (1982). The effect of exercise on amino acid oxidation and protein synthesis in sheep. In 'Energy Metabolism of Farm Animals'. European Assoc. for Animal Production EAAP Publication no. 29. (Eds A. Edern and Sundstol.) pp. 124-127. (The Agricultural University of Norway: Lillehammer, Norway.) Commonwealth Agricultural Bureaux (1980). 'The Nutrient Requirements of Ruminant Livestock.' K. Blaxter (Chair.) (Commonwealth Agricultural Bureaux: Slough, U.K.) Czok, R., and Lamprecht, W. (1974). Pyruvate, phophenolypyruvate and D-glycerate-2phosphate. In 'Methods of Enzymatic Analysis'. 2nd Edn, vol 3. (Ed. H. U. Bergmeyer.) pp. 1446-51. (Verlag Chemie: Weinheim.) Dunshea, F. R., Bell, A. W., and Trigg, T. E. (1990). Non-esterified fatty acid and glycerol kinetics and fatty acid re-esterification in goats during early lactation. B. J. Nutr. 64, 133-45. Faichney, G. J., Barker, P. J., Setchell, B. P., and Lindsay, D. B. (1981). The utilization of lactic acid by sheep in late pregnancy. Q. J. Exp. Physiol. 66, 195-201. Fisher, D. J., Heymann, M. A., and Rudolph, A. M. (1980). Myocardial oxygen and carbohydrate consumption in fetal lambs in utero and in adult sheep. Am. J. Physiol. 238, H399-H405. Garland, P. B., and Randle, P. J. (1962). A rapid enzymic assay for glycerol. Nature 196, 987-8. Harman, N. G., and Pethick, D. W., (1987). Metabolism of some gluconeogenic substrates in sheep during exercise. Proc. Nutr. Soc. Aust. 12, 35.


Herdt, T. H., Wensing, T., Haagsman, H. P., van Golde, L. M. G., and Brenkink, H. J. (1988). Hepatic triacylglycerol synthesis during a period of fatty liver development in sheep. J. Anim. Sci. 66, 1997-2013. Hickson, R. C., Rennie, M. J., Conlee, R. K., Winder, W. W., and Holloszy, J. 0 . (1977). Effects of increased plasma fatty acids o n glycogen utilization and endurance. J. Appl. Physiol. 43, 829-33. Jarrett, I. G., Filsell, 0. H., and Ballard, F. J. (1976). Utilisation of oxidizable substrates by the sheep hindlimb: effects of starvation and exercise. Metabolism 25, 523-31. Judson, G. J., Filsell, 0. H., and Jarrett, 1. G. (1976). Glucose and acetate metabolism in sheep a t rest and during exercise. Aust. J. Biol. Sci. 29, 215-22. Katz, M. L., and Bergman, E. N. ( 1 9 6 9 ~ )A . method for simultaneous cannulation of the major splanchnic vessels in the sheep. Am. J. Vet. Res. 30, 655-61. Katz, M. L., and Bergman, E. N. (196917). Simultaneous measurement of hepatic and portal venous bloodflow in the sheep and dog. Am. J. Physiol. 216, 946-52. Katz, A., Sahlin, K., and Henricksson, J. (1986). Muscle ammonia metabolism during isometric contraction in humans. Am. J. Physiol. 250, '334-40. Kjaer, M., Secher, N. H., Bach, F., and Galbo, H. (1987). Role of motor centre activity for homonal changes and substrate mobilization in exercising humans. Am. J. Physiol. 253, R687-R695. Lehmann, M., Keul, J., Huber, G., and Prada, D. Da. (1981). Plasma catechlamines in trained and untrained volunteers during graded exercise. Int. J. Sports Medicine 2 , 143-7. Lindsay, D. B. (1982). Relationships between amino acid catabolism and protein anabolism in the ruminant. Fed. Proc. 41, 2550-4. Lynch, J. J. (1967). Ranging behaviour of sheep in large areas. Proc. Aust. Ecol. Soc. 2, 167-9. Lynch, J. J. (1974). Behaviour of sheep and cattle in the more arid areas of Australia. In 'Studies of Australian Arid Zones-Pt 2: Animal Production. (Ed. A. D. Wilson.) pp. 37-49. (CSIRO: Melbourne.) McCandless, E. L., Woodward, B. A., and Dye, J. A. (1948). Alloxan diabetes in sheep under fasting and non-fasting conditions. Am. J. Physiol. 154, 94-106. McVeigh, J. M., Tarrant, P. V., and Harrington, M. G. (1982). Behavioral stress and skeletal muscle glycogen metabolism in young bulls. J. Anim. Sci. 54, 790-5. Mottram, R. F., Lynch, P. R., and Owen, 0. (1973). Forearm angiography during sustained isometric handgrip contractions. Invest. Radiol. 8, 22-7. Naylor, J. M., Kronfeld, D. S., Freeman, D. E., and Richardson, D.(1984). The saturable nature of hepatic lactate metabolism in sheep. Can. J. Anim. Sci. 6 4 (Suppl.), 271-2. Newsholme, E. A., and Leech, A. R. (1983). 'Biochemistry for the Medical Sciences.' Uohn Wiley & Sons: New York.) Noll, F. (1985). L-(+)-lactate determination. In 'Methods of Enzymatic Analysis'. 3rd Edn, Vol. 8. (Ed. H. U. Bergmeyer.) pp. 582-88. (VCH Verlagsgesellschaft: Weinheim.) Palmer, S. M., Oakes, G. K., Champion, J. A., Fisher, D. A., and Hobel, C. K. (1984). Catacholamine physiology in the ovine fetus: (111) maternal and fetal responses to acute maternal exercise. Am. J. Obstet. Gynecol. 149, 426-34. Pethick, D. W., Harman, N. G., and Chong, J. K. (1987). Non-esterified long-chain fatty acid metabolism in fed sheep at rest and during exercise. Aust. J. Biol. Sci. 40, 221-34. Pethick, D. W., and Harman, N. G. (1989). Limitations to endurance exercise in sheep. In 'Proceedings of 7th International Conference on Production Disease in Farm Animals'. pp. 61-64. July 1989. (Cornell Uni.: Ithaca, U.S.A.) Rowell, L. B. (1974). Human cardiovascular adjustments t o exercise and thermal stress. Physiol. Rev. 51, 75-159. Sahlin, K. (1986). Metabolic changes limiting muscle performance. In 'Biochemistry of Exercise'. International Series o n Sport Sciences, vol. 16. (Ed. B. Saltin.) pp. 323-43. (Human Kinetics Publ.: Champaign, Illinois.) Shirley, J. E., Emery, R. S., Convey, E. M., and Oxender, W. D. (1973). Enzymatic changes in bovine adipose tissue, semen and mammary tissue hormonal changes with initiation of lactation. J. Dairy Sc. 56, 569-4. Sonne, B., Mikines, K. J., and Galbo, H. (1987). Glucose turnover in 48-hour-fasted running rats. Am. J. Physiol. 252, R587-93.


Squires, V. R., and Wilson, A. D. (1971). Distance between food and water supply and its effect on drinking frequency, and food and water intake of Merino and Border Leicester sheep. Aust. J. Agric. Res. 22, 283-90. Squires, V. R., Wilson, A. D., and Daws, G. T. (1972). Comparisons of the walking activity of some Australian sheep. Proc. Aust. Soc. Anim. Prod. 9 , 376-80. Squires, V. R. (1974). Grazing distribution and activity patterns of Merino sheep on a saltbush community in south-east Australia. Appl. An. Ethology 1 , 17-30. Squires, V. R., (1976). Walking, watering and grazing behaviour of Merino sheep on two semi-arid rangelands in south-west New South Wales. Aust. Rangeland J. 1 , 13-23. Van der Walt, J. G., Baird, G. D. and Berman, E. N. (1983). Tissue glucose and lactate metabolism and interconversions in pregnant and lactating sheep. Br. J. Nutr. 50, 267-80. Vissing, J., Sonne, B., and Galbo, H. (1988). Role of metabolic feedback regulation in glucose production of running rats. Am. J. Physiol. 255, R400-6. Wasserman, K., Whipp, B. J., Koyal, S. N., and Beaver, W. L. (1973). Anaerobic threshold and respiratory gas exchange during exercise. J. Appl. Physiol. 3 5 , 236-43. Wasserman, K. (1986). The anaerobic threshold: definition, physiological significance and identification. Adv. Cardiol. 35, 1-23. Weekes, T. E. C. (1979). Carbohydrate Metabolism. In 'Digestive Physiology and Nutrition of Ruminants'. Vol. 2, Nutrition 2nd Edn. (Ed. D. C. Church.) pp. 187-209. (Oxford Press: Oregon.) Winder, W. W., Arogyasami, J., Yang, H. T., Thompson, K. G., Nelson, L. A., Kelly, K. P., and Han, D. H. (1988). Effects of glucose infusion in exercising rats. J. Appl. Physiol. 64, 2300-5. Zeirler, K. L. (1961). Theory of the use of arteriovenous concentration differences for measuring metabolism in steady and non-steady states. J. Clin. Invest. 40, 2111-25.

Manuscript received 3 May 1990, accepted 25 January 1991