The majority of teleost fish rely on venous blood for oxygen and nutrient supply to the heart. In these animals the myocardial tissue has a spongy,trabeculated architecture, with specific biochemical and mechanical characteristics (Rodnick and Williams,1999; Ewart and Driedzic,1987; Ewart et al.,1988; Tota, 1983; Agnisola and Tota, 1994). However, one third of teleost species, including salmonids, scombrids and anguillids, possess an outer layer of compact myocardium in the ventricle. This layer requires a blood supply via a dedicated vascular tree, the so-called coronary system, which usually originates from the branchial efferent, bringing arterious blood to the heart(Farrell and Jones, 1992).

As only part of the myocardial tissue is supplied by coronary arteries, it is likely that the fish heart does not have an obligatory dependence on its coronary blood flow (Farrell,2002). Only the skipjack tuna heart, whose ventricle contains up to 60% of compacta, is thought to be obligatorily dependent on its coronary circulation (Farrell et al.,1992). On the other hand, there are several indications that coronary circulation is essential under conditions of both hypoxia and intense swimming activity. In trout, coronary flow can increase during exercise or hypoxia (Gamperl et al.,1994), and coronary ligation resulted in a reduced proportion of compact myocardium together with a decrease in a series of metabolic enzymes(Farrell et al., 1990; Gamperl et al., 1994). Moreover, the long-term significance of coronary circulation was demonstrated by the fact that arteries can grow around the ablation site in 14 days(Farrell et al., 1990). The modulation of coronary flow under conditions of hypoxia or swimming suggests a neurohumoral and/or local regulation of the coronary resistance and consequently of the perfusion of the myocardium. Although the mechanism for such regulation in vivo is not known, a number of studies in vitro suggest that a complex interaction between various circulating and paracrine factors is involved (Mustafa et al., 1992; Mustafa and Agnisola, 1994; Agnisola et al., 1996; Mustafa and Agnisola, 1998).

There is no direct evidence for the mechanical significance of compacta perfusion. Information on the interplay between coronary physiology and ventricle performance in teleosts is derived from in situ and in vitro work. One limitation of these studies is that in all the preparations used, the heart was working against a fixed pressure (pressure head) rather than against a resistance. In vivo, aortic pressure and flow result from the matching of vascular resistance and the capacity of the heart to produce force and then to move blood. This cannot be simulated under pressure-head in vitro conditions, where the heart is constrained by the need to produce a minimal constant pressure in order to get flow. Also,coronary flow can barely be synchronised with heart requirements, something that in vivo is self-accommodated via changes in the pressure generated by the heart and local, paracrine and/or metabolic mechanisms. The aim of the present study was to define the role played by a well supplied compacta in fish heart performance using a simple protocol (with or without coronary perfusion) in an in vitro preparation in which the isolated trout heart was working against a resistance, as in vivo, and with the coronary flow related to the ventricular function on a beat-to-beat basis. The results provided evidence for some self-regulating properties of the teleost heart.

### Animals

Rainbow trout Oncorhynchus mykiss Walbaum of both sexes were obtained from a local fish farm (Randbøldal, Denmark). Fish were held in 400 liter tanks with running tapwater at 12±0.5°C under an artificial 12 h:12 h L:D photoperiod. They were fed 3 times a week with commercial trout pellets and allowed to acclimatize for 10 days prior to experiments. A batch of 29 animals was used for this study. The mean mass of the animals was 0.375±0.011 kg (± s.e.m.). The animals were divided into two groups: +CF, used to obtain an isolated heart preparation with perfusion of the coronary system (mean animal mass 0.362±0.015 kg; mean ventricle mass 0.297±0.011 g), and -CF,used to obtain an isolated heart preparation without perfusion of the coronary system (mean animal mass 0.393±0.018 kg; mean ventricle mass 0.327±0.017 g). There was no significant difference in animal or ventricle mass between the two groups.

### Dissection and preparation of the isolated trout heart

Fish were injected with heparin (∼1 ml kg-1 of a 0.9% NaCl solution containing 300 units heparin ml-1) in the caudal vein and killed by a quick blow on the head. The thorax was cut open and after removing the pericardium the entire heart was dissected out. The atrium, ventral aorta and coronary artery were cannulated as previously described(Mustafa et al., 1992; Agnisola et al., 1994). The isolated heart was pre-perfused with ice-cold saline solution during the cannulation procedure to clear the blood from the lumen and coronary arteries. The entire procedure from opening of thorax to coronary cannulation took approximately 15 min. After cannulation the heart was mounted in the perfusion chamber.

### The experimental set-up

The perfusion chamber, a double-jacketed chamber similar to that previously described (Houlihan et al.,1988; Agnisola et al.,1994), was connected to a water bath (RM6 Lauda, DRR.WOBSER GMBH& Co. KG, Lauda-Königshofen, Germany), which maintained a chamber temperature of 10±0.5°C. The input and output reservoirs were at the fixed heights of 7 cm and 10 cm from the level of saline in the chamber,respectively, and were connected to each other via a small tube as described by Agnisola et al.(1998)(Fig. 1), which enabled perfusion of the isolated heart with 50 ml of recirculating saline. The input and output loadings on the heart were set through resistances according to Agnisola et al. (1998). Briefly, the input loading on the heart was set by a variable resistance(Ri) and the saline level difference between the input reservoir and the heart. Output resistance was set by a fixed resistance(Ro) in the output tube (a teflon tube 17.8 cm long, 0.5 mm i.d.). The input to the coronaries was via a side arm on the output tube before Ro, and a constant resistance(Rc, consisting of a nylon tube 10 cm long, 0.3 mm i.d.)allowed the coronary input pressure to be directly related to the pressure generated by the heart. The experimental set-up was placed in a refrigerated cabinet set at 10°C. No detectable temperature changes occurred while opening and closing the glass door of the cabinet.

Fig. 1.

Scheme of the perfusion apparatus and pressure recording set-up. The heart was working against a resistance, resulting from the combined effects of Ro (= 6 TPa s m-3) and Rc(= 30 TPa s m-3) (i.e. the reciprocal of the sum of the conductances of the aortic tube and the coronary tube); however, as Ro>>Rc, afterload was depending mainly from Ro. The variable resistance Ri allowed us to set input pressure and flow. Input and output reservoirs were set at 7 and 10 cm, respectively, above the level of saline in the perfusion chamber. Pressures from the chamber (1), the atrium(2), and from saline filled tubes placed upstream (3) and downstream (4) of Ro, and downstream Rc (5), were sequentially acquired via a set of computer controlled valves connected with a pressure transducer. Inset: typical pressure trace recorded upstream Ro (3) under basal perfusion conditions.

Fig. 1.

Scheme of the perfusion apparatus and pressure recording set-up. The heart was working against a resistance, resulting from the combined effects of Ro (= 6 TPa s m-3) and Rc(= 30 TPa s m-3) (i.e. the reciprocal of the sum of the conductances of the aortic tube and the coronary tube); however, as Ro>>Rc, afterload was depending mainly from Ro. The variable resistance Ri allowed us to set input pressure and flow. Input and output reservoirs were set at 7 and 10 cm, respectively, above the level of saline in the perfusion chamber. Pressures from the chamber (1), the atrium(2), and from saline filled tubes placed upstream (3) and downstream (4) of Ro, and downstream Rc (5), were sequentially acquired via a set of computer controlled valves connected with a pressure transducer. Inset: typical pressure trace recorded upstream Ro (3) under basal perfusion conditions.

Once mounted in the perfusion chamber, the heart was wrapped with a transparent plastic film. Although not properly simulating a pericardium (no negative preload was generated), this procedure avoided overstretching of the atrium during volume loading of the preparation.

The saline in the input and output reservoirs was gassed with 99.5%O2 and 0.5% CO2 throughout the entire experiment, taking care not to trap any air bubbles in the system. The perfusate was Cortland saline (Wolf, 1963) as modified by Farrell et al.(1986) and Farrell(1987), of composition (in g l-1): NaCl, 7.25; KCl, 0.23,MgSO4·7H2O, 0.23;NaH2PO4·H2O, 0.016;NaHPO4·2H2O, 0.28; glucose, 1.0;polyvinylpyrrolidone (PVP) 10.0; CaCl2. 2H2O, 0.37; the pH was adjusted to 7.9 at 10°C with NaHCO3 (approx. 1 g l-1).

### Measurements and calculations

Atrial pressure was measured through a small saline-filled rigid tube, the side arm of the atrial-cannula support in the chamber wall, connected with the pressure transducer (Baxter, USA). Aortic and coronary pressures were measured through saline-filled side arms connected to the pressure transducer. Pressures were acquired sequentially at 5 min intervals from chamber, atrium and saline-filled tubes placed upstream and downstream of Ro and downstream of Rc, viaa computer controlled set of valves (Fig. 1). The inset in Fig. 1 is a typical pressure trace recorded upstream of Ro (aortic pressure), and shows that the pressure pulse was very similar to that usually recorded in vivo(Holeton and Randall, 1967; Wood and Shelton, 1980). Pressures (kPa) was measured using the pressure in the perfusion chamber as reference. Atrial preload (end-diastolic atrial pressure, Pa) and ventricular preload (mean atrial pressure, Pv) were determined from the atrial pressure recordings. Corrections were made for aortic and coronary cannulae resistances to obtain afterload (= mean aortic pressure) and coronary input pressure (= mean pressure in the coronary artery, Pc), respectively. At the end of each experiment, the ventricle was dissected, blotted dry and weighed. Heart rate(fh; beats min-1) was determined from pressure records. Cardiac output (; ml min-1 kg-1) and coronary flow (Fc;ml min-1 kg-1) were calculated from the fall of mean pressure (calculated from the pressure data acquired over a period of 10 s)through Ro and Rc, respectively. Stroke volume (Vs; ml kg-1) was determined by the relationship: /fh. The power output (PO; mW g-1) of the heart was calculated as:[(afterload - preload) (kPa) ×(+Fc) (ml min-1 kg-1) × animal mass (MA)/ ventricle mass (MV) / 60. Stroke work(WS; mJ g-1) was calculated as: PO×60/fh. Coronary resistance[Rc; TPa s m-3 (note that T=tera=1012)] was calculated as: Pc(kPa)×0.06/FC (ml min-1). The factor 0.06 is necessary to convert coronary flow to ml s-1 and pressure to TPa(Mustafa and Agnisola,1998).

Oxygen consumption of the heart was determined as follows:

$${\dot{V}}_{\mathrm{O}_{2}}=(P\mathrm{i}_{\mathrm{O}_{2}}-P\mathrm{o}_{\mathrm{O}_{2}}){\times}{\alpha}_{\mathrm{w},\mathrm{O}_{2}}{\times}{\dot{Q}}$$
⁠,where PiO2 is the input PO2, PoO2 is the output PO2, αw,O2is the oxygen solubility of the saline (ml O2 l-1g-1), and is the cardiac output in ml min-1 g-1. Oxygen partial pressure(PO2) was measured using a 16-730 oxygen electrode (Microelectrodes, Inc., Bedford, NH, USA) thermostatted at 10°C. The sensor was calibrated with sodium sulphite/borax solution for zero PO2, and with air-equilibrated saline for 20.94% oxygen. Linearity of the sensor response up to 100% oxygen was checked,and the percentage properly converted to oxygen partial pressure (in mmHg). Mechanical efficiency of the heart (i.e. the ratio between mechanical work and energy consumption, expressed as %) was determined as: 100 ×(PO×0.0498) × O2, where PO is the power output in mW and O2 is the rate of oxygen consumption in ml min-1 [assuming that 1 mW s-1 (1 mJ) is equivalent to 0.0498 μl O2min-1 (Davie and Franklin,1992)]. We also assumed that the heart mainly worked aerobically,so that the energy source of the heart was considered proportional to the oxygen consumption.

### Experimental protocols

#### Basal conditions

After mounting the isolated cannulated trout heart in the chamber, and putting an input load on the atrium, the heart started working against the fixed resistance, recirculating the 50 ml saline solution. Any leaks in the heart were discovered by a decrease in the saline level in the input reservoir and these hearts were discarded. The isolated working heart was left for a period of at least 30 min to stabilize, i.e. when the heart had settled into regular beating. Hearts that did not stabilize were discarded. At the end of the stabilization period, the cardiac output was adjusted to approximately 15 ml min-1 kg-1(Houlihan et al., 1988; Agnisola et al., 1994) by changing the variable input resistance (Ri). As soon as a basal cardiac output was obtained, automatic recording of pressure at 5 min intervals was begun; seven determinations were made for a total period of 30 min perfusion under basal conditions. At the end of this period, the oxygen consumption of the heart was determined.

At the end of the 30 min period of perfusion under basal conditions, the filling pressure was increased (by varying Ri) in seven steps of 5 min each and the pressure measured at each step. When maximal stroke work was reached, oxygen consumption was measured again.

### Statistics

All values are means ± s.e.m. One-way analysis of variance (ANOVA) was used to test the time course of the effect of different parameters during basal perfusion. Two-way ANOVA was used to analyse the changes in stroke work, ventricular preload and afterload during volume loading. P<0.05 was taken to indicate statistical significance. Statistics were performed with GraphPad Prism (GraphPad Software, Inc., San Diego, CA, USA).

### Effect of coronary perfusion on basal atrial performance

In the +CF group of hearts, the initial atrial preload necessary to get a cardiac output of about 15 ml min-1 kg-1 was 0.15±0.02 kPa (Fig. 2). Once set, this parameter did not change significantly during the successive 30 min of perfusion. At this load, the atrial performance remained constant with a mean atrial pressure (ventricular preload) that did not change from the initial value of 0.26±0.02 kPa. Fig. 3 shows that the ventricular performance parameters also remained constant at the initial level throughout the 30 min of basal perfusion.

Fig. 2.

Atrial (squares) and ventricular (circles) preload values during 30 min perfusion of the isolated working trout heart under basal conditions, with(+CF, open symbols) and without (-CF, closed symbols) coronary perfusion. At time 0 the atrial preload was set in order to get a cardiac output of about 15 ml min-1 kg-1. Any successive change indicates a self-adjustment of the heart. Two-way ANOVA indicated a significant difference between +CF and -CF data (P<0.01) in both parameters. Repeated-measures one-way ANOVA and Dunnet post-hoc test applied to-CF data indicated a significant increase in both atrial and ventricular preload 5 min after the beginning of perfusion (P<0.05).

Fig. 2.

Atrial (squares) and ventricular (circles) preload values during 30 min perfusion of the isolated working trout heart under basal conditions, with(+CF, open symbols) and without (-CF, closed symbols) coronary perfusion. At time 0 the atrial preload was set in order to get a cardiac output of about 15 ml min-1 kg-1. Any successive change indicates a self-adjustment of the heart. Two-way ANOVA indicated a significant difference between +CF and -CF data (P<0.01) in both parameters. Repeated-measures one-way ANOVA and Dunnet post-hoc test applied to-CF data indicated a significant increase in both atrial and ventricular preload 5 min after the beginning of perfusion (P<0.05).

Fig. 3.

Ventricular performance during 30 min perfusion of the isolated working trout heart under basal conditions, with (+CF, open symbols) and without (-CF,closed symbols) coronary perfusion. (A) Cardiac output (circles) and power output (squares); (B) stroke volume (circles) and stroke work (squares); (C)afterload (circles) and heart rate (squares). Two-way ANOVA indicated a significant difference between +CF and -CF data (P<0.01) for cardiac output and power output only.

Fig. 3.

Ventricular performance during 30 min perfusion of the isolated working trout heart under basal conditions, with (+CF, open symbols) and without (-CF,closed symbols) coronary perfusion. (A) Cardiac output (circles) and power output (squares); (B) stroke volume (circles) and stroke work (squares); (C)afterload (circles) and heart rate (squares). Two-way ANOVA indicated a significant difference between +CF and -CF data (P<0.01) for cardiac output and power output only.

In the -CF group, although the initial set of atrial preload was not significantly different from that of the +CF group, the atrium adjusted within 5 min to a significantly higher preload value(Fig. 2), which remained constant during the remaining perfusion period at the basal setting. As a consequence, mean atrial pressure also increased up to about 0.3 kPa. The increased preload in the -CF group of hearts was associated with a significant reduction in cardiac output and power output(Fig. 3A). Stroke volume and stroke work were consistently lower in the -CF hearts(Fig. 3B), although the difference was not statistically significant. There was no difference in afterload and heart rate between the two groups(Fig. 3C).

In the +CF group, coronary performance, as evaluated by coronary pressure,flow and resistance values, was constant throughout the 30 min perfusion period under basal conditions (Fig. 4).

Fig. 4.

Coronary flow (filled circles), pressure (open circles) and resistance(open squares) during 30 min perfusion of the isolated working trout heart under basal conditions in the +CF group of hearts. One-way ANOVA indicated that these three parameters were not dependent on perfusion time.

Fig. 4.

Coronary flow (filled circles), pressure (open circles) and resistance(open squares) during 30 min perfusion of the isolated working trout heart under basal conditions in the +CF group of hearts. One-way ANOVA indicated that these three parameters were not dependent on perfusion time.

### Effect of coronary perfusion on the heart's response to volume loading

Fig. 5.

Volume loading of the isolated working trout heart with coronary perfusion included (+CF, open circles) or excluded (-CF, filled circles). Two-way ANOVA indicated a significant effect of both coronary perfusion (P<0.01)and preload (P<0.001) on stroke work. (Inset A) Relationship between coronary flow and preload in the +CF group. One-way ANOVA indicated no-significant effect of preload on coronary flow. (Inset B) Effect of ventricular preload on afterload of the isolated working trout heart with coronary perfusion included (+CF, open circles) or excluded (-CF, filled circles). Repeated-measures one-way ANOVA indicated a significant effect of preload on afterload in both groups. Two-way ANOVA indicated no significant effect of coronary perfusion on afterload.

Fig. 5.

Volume loading of the isolated working trout heart with coronary perfusion included (+CF, open circles) or excluded (-CF, filled circles). Two-way ANOVA indicated a significant effect of both coronary perfusion (P<0.01)and preload (P<0.001) on stroke work. (Inset A) Relationship between coronary flow and preload in the +CF group. One-way ANOVA indicated no-significant effect of preload on coronary flow. (Inset B) Effect of ventricular preload on afterload of the isolated working trout heart with coronary perfusion included (+CF, open circles) or excluded (-CF, filled circles). Repeated-measures one-way ANOVA indicated a significant effect of preload on afterload in both groups. Two-way ANOVA indicated no significant effect of coronary perfusion on afterload.

### Effect of coronary perfusion on oxygen consumption and mechanical efficiency of the trout heart.

The total oxygen consumption and the mechanical efficiency of the heart measured under basal perfusion conditions and at maximal volume loading are shown in Fig. 6. In both +CF and -CF hearts, volume loading induced a significant increase in rate of oxygen consumption, from a basal level of approximately 20 μl O2min-1 g-1 to approx. 40 μl O2min-1 g-1 at maximal volume loading. In both groups,basal mechanical efficiency was about 17% and was not affected by volume loading. Neither rate of oxygen consumption nor efficiency were affected by coronary perfusion.

Fig. 6.

Effect of coronary perfusion on (A) oxygen consumption and (B) mechanical efficiency of the isolated working trout heart under basal perfusion conditions and at maximal volume loading. Two-way ANOVA with Bonferroni post-hoc test indicated a significant effect of preload on O2, but no significant difference was observed between +CF and -CF. Efficiency was not affected by either the perfusion conditions or the presence of coronary perfusion.

Fig. 6.

Effect of coronary perfusion on (A) oxygen consumption and (B) mechanical efficiency of the isolated working trout heart under basal perfusion conditions and at maximal volume loading. Two-way ANOVA with Bonferroni post-hoc test indicated a significant effect of preload on O2, but no significant difference was observed between +CF and -CF. Efficiency was not affected by either the perfusion conditions or the presence of coronary perfusion.

The heart is an organ with complex self-regulating properties, in which the interaction between myocytes, coronary vasculature and neurohumoral factors play a central role (Kresh and Armour,1997). The self-regulating mechanisms of the heart are tuned to provide a cardiac output that is constantly adequate for the demand, and are constrained by the continuous need to match heart function with the compliance and resistance properties of arterial vessels(Berne and Levy, 1992).

A main feature of the present study is the use of a resistance-headed isolated heart preparation. In this preparation, the actual aortic pressure and flow result from matching a fixed resistance (Ro),simulating the peripheral resistance in vivo, and the intrinsic capacity of the heart to produce force and then to move the perfusion saline(Agnisola et al., 1998). The resistance-headed condition, as occurs in vivo, was hereby reproduced. More importantly, in this set-up the coronary flow was continuously and automatically linked to the pressure generated by the heart. In vivo, ventral aorta pressure and gill resistance determine the driving force for coronary perfusion (the coronary pressure), a situation that was reproduced in the perfusion set-up we used thanks to the fixed resistance Rc. This allowed maintenance of the intrinsic and reciprocal interactions between ventricle and coronaries on a beat-to-beat basis, as occurs in vivo. While lacking the potential to dissect the role of single specific features on heart performance (e.g. the effect of pressure loading), the resistance-headed preparation that we used did allow the isolated heart preparation to express some of the intrinsic self-regulating mechanisms that determine the heart's global pump function in vivo (Kresh and Armour,1997). Thus, this preparation is well suited for the study of the relative significance of a well-perfused compacta on trout heart performance.

The basal performance of the resistance-headed heart with the perfused coronary vascular tree (+CF group) was close to that of a similar preparation previously reported (Agnisola et al.,1998), and similar to that of in situ or in vitro pressure-headed preparations(Farrell et al., 1986; Houlihan et al., 1988; Davie and Farrell, 1991; Davie et al., 1992). As no attempt was made to simulate the role of the pericardium in determining absolute values of atrial preload (Farrell et al., 1988), the input pressure necessary to get basal cardiac output was higher than the ambient pressure (approx. 0.1 kPa), and similar to that reported previously for both pressure-headed(Houlihan et al., 1988) and resistance-headed (Agnisola et al.,1998) isolated trout heart preparations. The pressure generated by the heart was similar to that reported in vivo in the trout ventral aorta (Kiceniuk and Jones,1977; Wood and Shelton,1980), while the coronary pressure was somewhat lower than the in vivo dorsal aorta pressure(Kiceniuk and Jones, 1977; Gamperl et al., 1995). This was probably a consequence of the lower coronary resistance in vitrowith respect to the value in vivo, due to the lower viscosity of saline with respect to blood (by a factor of 2.5-3.5; Farrell, 1987). In agreement with previously reported data in vitro(Agnisola et al., 1998; Houlihan et al., 1988),coronary flow was 2-5% of cardiac output, and higher than the in vivovalues (1-2%, Axelsson and Farrell,1993; Gamperl et al.,1994), probably because both the viscosity and oxygen content of saline were low compared with blood.

The absence of a coronary supply did not affect the heart rate and pressure generated by the heart, but significantly did affect the heart's capacity to maintain basal cardiac output, initially set at 15 ml min-1kg-1, resulting in a significant reduction in the heart's power output. This result suggests that coronary perfusion was significantly affecting cardiac contractility. However, the consequences of this effect were apparently limited by self-regulating mechanisms, which led to a significant increase in preload which, via the Starling mechanism, would oppose the stroke volume reduction. This may explain the non-significant, although consistent, reduction in stroke volume and stroke work.

The dependence of cardiac contractility on a well-supplied compacta was confirmed by the response of hearts to the volume-loading protocol. This was accomplished by a stepwise increase in input pressure, with no attempt to control other parameters. In particular, as the heart was working against a fixed resistance, the increase in atrial preload caused a significant increase in both stroke volume and afterload (Berne and Levy, 1992). Thus the maximal stroke work thereby obtained was the combination of an increase in both pressure and volume work. In the +CF hearts, this effect was probably part of intrinsic adjustments that resulted in a constant coronary flow despite the increased coronary resistance due to the increased extravascular compression produced by the myocardium (systolic strangulation of coronary perfusion; Gamperl et al., 1995). In the absence of coronary perfusion, the shift of the stroke work-preload curve towards the right, shown in Fig. 5, confirmed that a main effect of coronary perfusion and a main role for the compacta were the increase in ventricular contractility, allowing the heart to produce a greater stroke work at the same preload.

Several theoretical models based on the mammalian left ventricle have suggested that the vascular and ventricular properties are matched to achieve maximal transfer of mechanical energy or maximal ventricular metabolic efficiency (Sunagawa et al.,1983; Burkhoff and Sagawa,1986; Van den Horn et al.,1985; Toorop et al.,1988). De Tombe et al.(1993) have suggested that,when confronted with variable vascular loading, these parameters are both kept nearly maximal, although not precisely optimised. An interesting result of the present study is the apparent independence of mechanical efficiency from volume loading. This contrasts with previous data, obtained using pressure-headed fish heart preparations(Farrell and Steffensen, 1987; Davie and Franklin, 1992),showing lower efficiency at lower loads. It is possible that the combined effects of the increased stroke volume (which should increase efficiency; Shipke, 1994) and afterload(which may decrease efficiency; Shipke,1994) would help to maintain efficiency at a constant, nearly optimal level. On the other hand, neither oxygen consumption nor efficiency were affected by coronary perfusion. This result, which contrasts with the reduction observed by Houlihan et al.,(1988) on a pressure-headed preparation, may in part reflect the fact that coronary perfusion did not affect afterload and heart rate, two major determinants of oxygen consumption in fish (Farrell and Jones,1992). However, it may also indicate that the interplay and integration of the different homeodynamic mechanisms operating in the heart,help control its energetics under a wide range of variations of extrinsic and intrinsic factors, including coronary perfusion.

This work was supported by funds from the Danish Natural Sciences Research Council. The authors thank Bent Bach Andersen for developing the computer-operated pressure and flow recording system and Lilian Salling for the technical help during the course of the study.

Agnisola, C., Venzi, R., Houlihan, D. and Tota, B.(
1994
). Coronary flow-pressure relationship in the working isolated fish heart: trout (Oncorhynchus mykiss) versus torpedo(Torpedo marmorata).
Phil. Trans. R. Soc. Lond. B
343
,
189
-198.
Agnisola, C., Mustafa, T. and Hansen, J. K.(
1996
). Autoregulatory index, adrenergic responses and interaction between adrenoreceptors and prostacyclin in the coronary system of rainbow trout.
J. Exp. Zool.
275
,
239
-248.
Agnisola, C. and Tota, B. (
1994
). Structure and function of the fish ventricle: flexibility and limitations.
Cardiosci.
5
,
145
-154.
Agnisola, C., Jensen, F. B., Tota, B. and Mustafa, T.(
1998
). Performance of the isolated rainbow trout heart perfused under self-controlled coronary pressure conditions: effects of high and low oxygen tension, arachidonic acid and indomethacin.
J. Comp. Physiol. B
168
,
96
-104.
Axelsson, M. and Farrell, A. P. (
1993
). Coronary blood flow in vivo in the coho salmon (Oncorhynchus kisutch).
Am. J. Physiol.
264
,
R963
-R971.
Berne, R. M. and Levy, M. N. (
1992
).
Cardiovascular Physiology.
St Louis: Mosby-Year Book.
Burkhoff, D. and Sagawa, K. (
1986
). Ventricular efficiency predicted by an analytical model.
Am. J. Physiol.
250
,
R1021
-R1027.
Davie, P. S. and Farrell, A. P. (
1991
). Cardiac performance of an isolated heart preparation from the dogfish (Squalus acanthias): the effects of hypoxia and coronary artery perfusion.
Can. J. Zool.
69
,
1822
-1828.
Davie, P. S. and Franklin, G. E. (
1992
). Myocardial oxygen consumption and mechanical efficiency of a perfused dogfish heart preparation.
J. Comp. Physiol. B
162
,
256
-262.
Davie, P. S., Farrell, A. P. and Franklin, G. E.(
1992
). Cardiac performance of an isolated eel heart: Effects of hypoxia and responses to coronary artery perfusion.
J. Exp. Zool.
262
,
113
-121.
De Tombe, P. P., Jones, S., Burkhoff, D., Hunter, W. C. and Kass, D. A. (
1993
). Ventricular stroke work and efficiency both remain nearly optimal despite altered vascular loading.
Am. J. Physiol.
264
,
H1817
-H1824.
Ewart, H. S. and Driedzic, W. R. (
1987
). Enzymes of energy metabolism in salmonid hearts: spongy versus cortical myocardia.
Can. J. Zool.
65
,
623
-627.
Ewart, H. S., Canty, A. A. and Driedzic, W. R.(
1988
). Scaling of cardiac oxygen consumption and enzyme activity levels in sea raven (Hemitripterus americanus).
Physiol. Zool.
61
,
50
-56.
Farrell, A. P. (
1987
). Coronary flow in a perfused rainbow trout heart.
J. Exp. Biol.
129
,
107
-203.
Farrell, A. P. (
2002
). Coronary arteriosclerosis in salmon: growing old or growing fast?
Comp. Biochem. Physiol.
132A
,
723
-735.
Farrell, A. P. and Jones, D. R. (
1992
). The heart. In
Fish Physiology
, vol.
XIIA
(ed. W. S. Hoar, D. J. Randall and A. P. Farrell),pp.
1
Farrell, A. P. and Steffensen, J. F. (
1987
). An analysis of the energetic cost of the branchial and cardiac pumps during sustained swimming in trout.
Fish Physiol. Biochem.
4
,
73
-79.
Farrell, A. P., MacLeod, K. R. and Chancey, B.(
1986
). Intrinsic mechanical properties of the perfused rainbow trout heart and the effects of catecholamines and extracellular calcium under control and acidotic conditions.
J. Exp. Biol.
125
,
319
-345.
Farrell, A. P., Johansen, J. A. and Graham, M. S.(
1988
). The role of the pericardium in cardiac performance of the trout (Salmo gairdneri).
Physiol. Zool.
61
,
213
-221.
Farrell, A. P., Johansen, J. A., Steffensen, J. F., Moyes, C. D., West, T. G. and Suarez, R. K. (
1990
). Effects of exercise-training and coronary ablation on swimming performance, heart size and cardiac enzymes in rainbow trout (Oncorhynchus mykiss).
Can. J. Zool.
68
,
1174
-1179.
Farrell, A. P., Davie, P. S., Franklin, C. E., Johansen, J. A. and Brill, R. W. (
1992
). Cardiac physiology in tunas: I. In vitro perfused heart preparations from yellowfin and skipjack tunas.
Can. J. Zool.
70
,
1200
-1210.
Gamperl, A. K., Pinder, A. W. and Boutilier, R. G.(
1994
). Effect of coronary ablation and adrenergic stimulation on in vivo cardiac performance in trout (Oncorhynchus mykiss).
J. Exp. Biol.
186
,
127
-143.
Gamperl, A. K., Axelsson, M. and Farrell, A. P.(
1995
). Effects of swimming and environmental hypoxia on coronary blood flow in rainbow trout.
Am. J. Physiol.
269
,
R1258
-R1266.
Holeton, G. F. and Randall, D. J. (
1967
). The effect of hypoxia upon the partial pressure of gases in the blood and water afferent and efferent to the gills of rainbow trout.
J. Exp. Biol.
46
,
317
-327.
Houlihan, D. F., Agnisola, C., Lyndon, A. R., Gray, C. and Hamilton, N. M. (
1988
). Protein synthesis in a fish heart:responses to increased power output.
J. Exp. Biol.
137
,
565
-587.
Kiceniuk, J. W. and Jones, D. R. (
1977
). The oxygen transport system in trout (Salmo gairdneri) during sustained exercise.
J. Exp. Biol.
69
,
247
-260.
Kresh, J. Y. and Armour, J. A. (
1997
). The heart as a self-regulating system: integration of hemodynamic mechanisms.
Technol. Health Care
5
,
159
-169.
Mustafa, T. and Agnisola, C. (
1994
). Vasoactivity of prostanoids in the trout (Oncorhynchus mykiss)coronary system: modification by noradrenaline.
Fish Physiol. Biochem.
13
,
249
-261.
Mustafa, T. and Agnisola, C. (
1998
). Vasoactivity of adenosine in the trout (Oncorhynchus mykiss) coronary system: involvement of nitric oxide and interaction with noradrenaline.
J. Exp. Biol.
201
,
3075
-3083.
Mustafa, T., Agnisola, C. and Tota, B. (
1992
). Myocardial and coronary effects of exogenous arachidonic acid on the isolated and perfused heart preparation and its metabolism in the heart of trout(Oncorhynchus mykiss).
Comp. Biochem. Physiol.
103C
,
163
-167.
Rodnick, K. J. and Williams, S. R. (
1999
). Effects of body size on biochemical characteristics of trabecular cardiac muscle and plasma of rainbow trout (Oncorhynchus mykiss).
Comp. Biochem. Physiol.
122A
,
407
-413.
Shipke, J. D. (
1994
). Cardiac efficiency.
Basic Res. Cardiol.
89
,
207
-240.
Sunagawa, K., Maughan, W. L., Burkhoff, D. and Sagawa, K.(
1983
). Left ventricular interaction with arterial load studied in isolated canine ventricle.
Am. J. Physiol.
245
,
H773
-H780.
Toorop, G. P., Van den Horn, G. J., Elzinga, G. and Westerhof,N. (
1988
). Matching between feline left ventricle and arterial load: optimal external power and efficiency.
Am. J. Physiol.
254
,
H279
-H285.
Tota, B. (
1983
). Vascular and metabolic zonation in the ventricular myocardium of mammals and fishes.
Comp. Biochem. Physiol.
76A
,
423
-437.
Van den Horn, G. J., Westerhof, N. and Elzinga, G.(
1985
). Optimal power generation by the left ventricle. A study in the anesthetized open thorax cat.
Circ. Res.
56
,
252
-261.
Wolf, K. (
1963
). Physiological salines for fresh-water teleosts.
Prog. Fish. Cult.
25
,
135
-140.
Wood, C. M. and Shelton, G. (
1980
). Cardiovascular dynamics and adrenergic responses of the rainbow trout in vivo.
J. Exp. Biol.
87
,
247
-270.