Effects of ischaemia on vasculature

The effects of ischaemia on the structure and function of coronary vessels have been evaluated most thoroughly in the open-chest anaesthetised dog. In the beating heart the flow into the capillary bed is controlled by precapillary sphincters in the terminal arterioles. Short periods (up to 20 min) of ischaemia result during reperfusion in vasodilatation which is mediated through increased tissue adenosine levels. Longer periods of ischaemia (60 or more min) result in degenerative vascular changes and no-reflow during reperfusion. Endothelial cells swell, lose their pinocytotic vesicles and form spherical cytoplasmic protrusions into the capillary lumina. Prolonged ischaemia (over 3 hours) causes ruptures in microvascular walls. Such vessels become permeable to large tracers, e.g. carbon particles, and reperfusion results in haemorrhage. Capillaries in ischaemic myocardium are collapsed by compression resulting from intra- and extracellular oedema and contracture of myocytes. Erythrocytes plug collapsed microvessels. However, in vitro observations have shown that no-reflow may result from ischaemia even without erythrocyte plugging. Capillary compression is the main determinant of no-reflow after ischaemia. Degenerative ischaemic changes in vascular walls further contribute to vascular incompetence in vivo.

Morphological changes in endocardium subjected to global ischaemia

To provide insight into the effects of severe ischaemia on endocardium, the sequence of morphological changes which develop in the endocardium of the isolated rat heart subjected to 0-12 hours of global ischaemia at 37 degrees C was examined. A progression of changes occurred. Following one or more hours of ischaemia crater-like depressions and blebs appeared on the luminal surfaces of ventricular endothelial cells, with margination and clumping of nuclear chromatin, loss of glycogen granules, swelling of mitochondria, and the development of subendothelial membrane-bound dilatations of myocytes. Following two or more hours of ischaemia there was progressive separation of endothelial cells along their intercellular boundaries and desquamation of an increasing proportion of these cells. In regions of desquamation the surface was initially smooth due to persistence of the lamina densa of the basal lamina, but after longer periods of ischaemia the surface became rough with exposure first of the subendothelial connective tissue fibres (4 hours) and eventually of cardiac muscle cells (12 hours).

Scanning electron microscopy of heart muscle freeze-dried from dimethylsulfoxide for simultaneous demonstration of cell morphology and microvascular function

Scanning electron microscopy was used to examine cryofracture surfaces of ventricular myocardium from glutaraldehyde fixed rat and rabbit hearts subjected to intravascular injection of polymerizing acrylic resin. This allowed simultaneous observation of morphological features of cardiac muscle cells and the functional state of their associated small blood vessels. Because the resin injected to identify capillaries accessible to flow might be soluble in commonly used tissue dehydrating agents, alternative preparation methods using the cryoprotectants dimethylsulfoxide (DMSO) and glycerol were investigated. Provided a high performance backscattered electron detector and simple environmental cell were used to abolish specimen charging and circumvent potential instrument contamination, immersion in 2.82 M DMSO for 12 hr prior to cryofracture and freeze-drying gave the best results. The SEM appearance of specimens dehydrated in this way differed little from that of specimens prepared by ethanol dehydration and freeze-drying or by acetone dehydration and critical-point drying. Tissue shrinkage was 26.5 +/- 9.4%, comparable to that found after standard methods using solvent dehydration and critical-point drying.

Microvascular function at the margins of early experimental myocardial infarcts in isolated rabbit hearts

Injection of low-viscosity resin was used to identify in situ functional blood vessels at the margins of developing regional myocardial infarcts. The ventral interventricular branch (VIB) of the left coronary artery was occluded for 0-240 min in 20 isolated perfused rabbit hearts. After perfusion fixation with glutaraldehyde, resin was injected into the coronary arteries--that injected into the VIB contained dispersed lead dioxide and that injected into the remainder of the heart contained Fat Red 7B dye. This allowed macroscopic and microscopic identification of functional blood vessels. Following transmural freeze fracture, left ventricles were examined using back-scattered electron imaging in a scanning electron microscope. Close to 60% of capillaries in nonischemic myocardium allowed the passage of resin. Thirty minutes of ischemia produced a hyperemic increase to 80%-90% in the proportion of filled vessels. After 60 min, however, a severe reperfusion defect corresponding to the "no-reflow" phenomenon had developed, with virtually all vessels collapsed and less than 10% functional. Among the structurally normal myocytes adjacent to the infarct margin there was a significant reduction (to 30%-40%) in the proportion of functional capillaries. This was due to groups of dilated vessels which were not accessible to arterial supply. Although these marginal "low-flow" regions were of small volume at any one point in time, they seem likely to contribute to the progression of ischemic necrosis, and are probably nonfunctional due to the compression of their venous drainage traversing the infarct.

Location of bacteria in the mid-colon of the rat

The distribution of microorganisms in the mid-colon of the rat was studied by light and scanning electron microscopy. An antiserum against rat colon mucus was used to stabilize the mucus in situ. In samples not incubated with antiserum, the mucus disintegrated and contracted into patchy strands only partly covering the luminal surface of the colon. Bacteria were seen within fecal pellets, tangled among the strands of mucus, and scattered on the epithelial surface. However, when incubated with antiserum, mucus almost completely filled the lumen and coated the fecal pellets. Bacteria in these stabilized preparations were limited mainly to the fecal pellets, and there were small numbers scattered in the luminal mucus, but none were observed on the epithelial surface or within the crypts. Latex particles introduced into the lumen with the antiserum or with phosphate-buffered saline showed the same distribution as the bacteria. These findings are at variance with previous reports that organisms occur in abundance in the mucous layer, adjacent to cell surfaces, and inside crypts. Our results suggest that conventional preparation for microscopy without prior stabilization of the mucus in situ may lead to artifactual redistribution of microorganisms and emphasize the importance of mucus in maintaining mucosal-floral homeostasis in the colon.

The pathogenesis of stereocilia abnormalities in acoustic trauma

Stereocilia abnormalities develop in anesthetized guinea pigs within two minutes of exposure to loud sound (3 kHz, 125 dB SPL). Displacement, fracture, fusion and membrane rupture were observed well before the development (after 10 or more minutes) of fine structural changes to other cytoplasmic organelles within the hair cells. Thus stereocilia abnormalities can arise very rapidly due to the direct effects of noise on their microfilaments, cross linkages and limiting membranes. Such abnormalities are thus not necessarily secondary to changes in the body of the affected hair cells.

Graft detachment, a cause of incompetence in stent-mounted aortic valve allografts

The detachment of allograft tissue from supporting stent posts has been a common mode of failure of stent-mounted aortic allografts. In an effort to reduce the localized stress loading on the tissue at the top of the stent posts, two changes were introduced to the previous fabrication protocol followed by Green Lane Hospital. Specifically, they were the use of a flexible acetyl-copolymer stent and covering of the peripheral attachment sutures with a bias strip of Dacron cloth. This study showed that these changes did not reduce the incidence of allograft failure caused by graft detachment. Unexpectedly, covering antibiotic-treated allograft tissue with a bias strip actually increased the likelihood of failure of the valve, by accelerating the biological degradation of the underlying graft aortic wall. The weakened aortic wall tissue was consequently less able to resist the high stress loadings at the top of the stent posts where detachment first occurred.

The development and progression of myocyte injury at the margins of experimental myocardial infarcts

Distinct differences in the extent and progression of the lateral and epicardial boundaries of evolving regional infarcts were demonstrated in isolated rabbit hearts. Ischemia was produced by interrupting (0-240 minutes) flow in the ventral interventricular branch of the left coronary artery, whilst the remainder of the heart was continuously perfused with oxygenated Krebs-Henseleit bicarbonate buffer. Perfusion fixed blocks were freeze-fractured then examined using back-scattered electron imaging in a scanning electron microscope. Control myocytes showed relatively smooth, continuous internal fracture faces. After 30 min of ischemia myocytes showed evidence of mild, probably reversible, injury in the form of prominence of pits and channels. Severe injury, characterized by separation of organelles and prominent intracellular spaces, developed after 60 or more min of ischemia, first in the subendocardial two thirds, and after 120 min across the full thickness of the ventricular wall. At the lateral margins of infarcts there was a distinct cell-to-cell boundary between control and severely injured myocytes, with only a few scattered mildly injured cells within 30 mu of the infarct. Although transmural progression of necrosis provides the potential for recovery of the external aspect of the myocardium in the ischemic zone by reperfusion, corresponding regions of salvageable myocytes at lateral infarct margins are very narrow.

Reversal of the no reflow phenomenon in globally ischemic rat hearts by ventricular dilation

Changes in the contracture and stiffness of ventricular walls during ischemia and their effect on vascular reperfusion were studied in isolated rat hearts. Global ischemia was induced by stopping the flow of oxygenated perfusate for 60 min. Contracture pressures generated against water-filled left ventricular balloons maintained at diastolic volume increased between 8 and 25 min of ischemia and declined thereafter. On the other hand, left ventricular wall stiffness, estimated from the pressure required to periodically inflate otherwise uninflated balloons to diastolic volume, increased rapidly between 20 and 30 min of ischemia, and more slowly thereafter. Inflation of balloons following 30 or more min of ischemia caused significant reductions in left ventricular wall stiffness. Similar brief inflation after 60 min of ischemia allowed aqueous sodium fluorescein to perfuse the subendocardial half of the left ventricular wall which was otherwise not accessible to reperfusion. This reversal of the no reflow phenomenon was accompanied by a reduction in myocardial wall stiffness, stretching of myocytes and increased patency of myocardial capillaries.

Cardiac performance of isolated beating hearts obtained from rats anesthetized by three different agents

Two functional performance parameters (heart rate and coronary flow rate) of three groups of six isolated beating hearts obtained from rats anesthetized by inhalation of diethyl ether in air, injection intraperitoneally of sodium pentobarbital (60mg/kg) or intravenously of alphaxalone/alphadolone (15mg/kg) were compared. Differences were small, but alphaxalone-alphadolone showed lowest stable mean coronary flow rate and diethyl ether the widest variation between animals. No significant difference in function was found between ether and pentobarbital, currently the two most widely used methods. But, pentobarbital showed higher stable functional performance and least variation between animals. We conclude that pentobarbital is the most useful of these three agents for obtaining hearts for perfusion as isolated beating organs, and that alphaxalone-alphadolone is clearly less suitable.

Changing relationships between structure and function in the cochlea during recovery from intense sound exposure

This study examines the relationship between the morphological and electrophysiological changes in the guinea pig cochlea at intervals following exposure to intense sound (5 kHz, 125 dB SPL, 30 minutes). The cochlear compound action potential (N1) audiogram (2-30 kHz) was determined within one hour or 1, 7, 14, or 28 days after exposure. When these organs of Corti were examined by scanning electron microscopy, most (92%) showed damage to hair cells and supporting cells around the 5-kHz region. The number of damaged hair cells increased significantly (p less than 0.01) in the first 24 hours following exposure, but there was no significant change thereafter. In contrast, the initial extensive loss of function (maximum at 8-12 kHz) recovered partially during the first 24 hours but also showed no significant further change. All morphological lesions occurred within regions corresponding tonotopically to the N1 threshold losses, but only the larger lesions, which were observed seven or more days after exposure, fully reflected the extent of the loss of N1 thresholds. This indicates that functionally important damage to the cochlea is usually more extensive than indicated by topographical abnormalities in the organ of Corti.

The influence of the no-reflow phenomenon on reperfusion and reoxygenation damage and enzyme release from anoxic and ischaemic isolated rat hearts

Eighty isolated rat heart preparations were used to study relationships among creatine kinase (CK) release, the loss of vascular competence (no-reflow), and the distribution of morphological changes across the left ventricular wall which occur during 60 min global ischaemia or anoxia and following subsequent oxygenated reperfusion. Hearts were either fixed with glutaraldehyde for light and electron microscopy or were injected with 1% fluorescein to define the distribution of perfusable vessels. The extent of no-reflow in half of the hearts was reduced experimentally by maintaining the diastolic volume of the left ventricular lumen during ischaemia and anoxia with a water-filled balloon. The amount of CK released during 20 min of reoxygenation or reperfusion was inversely proportional to the extent of the no-reflow area observed just prior to reoxygeneration, and also reflected the transmural extent and the severity of myocardial cell damage. Extensive contraction band necrosis was only observed in reperfused regions of anoxic hearts. In isovolumic hearts reoxygenation caused no-reflow to develop in the ventricular myocardium, and this appeared to be associated with hypercontraction. Thus the no-reflow phenomenon has a profound effect on the transmural distribution of myocardial cell damage and enzyme release which follows post ischaemic reperfusion and post anoxic reoxygenation.

The accuracy of hair cell counts in determining distance and position along the organ of Corti

The relationship between the density of hair cells (cells/mm) and measured distance along the guinea pig organ of Corti was determined using light microscopy and the surface specimen technique. It was demonstrated that the density of inner hair cells (IHC; mean 92.0 +/- 2.4) and 1st row of outer hair cells (OHC1; mean 118.7 +/- 2.3) did not show significant variation along the organ of Corti except within 0.5-1.0 mm of the apex and base where there was considerable variation between animals in the density of cells. There was a close relationship between the accumulated number of either IHC or OHC1 and distance from the base along the organ of Corti. Distances estimated by hair cell counts were similar to those determined by direct measurement. It is concluded that hair cell counts can be used to reliably estimate distances along the organ of Corti where accurate direct measurement is not possible, such as in scanning electron microscopy.

The effect of coronary pressure on contracture and vascular perfusion in the hypoxic isolated rat heart

The relationships among coronary perfusion pressure (0-150 cm H2O), ventricular contracture, vascular resistance, and loss of perfusion of subendocardial myocardium were studied in anoxic isolated rat hearts. Coronary flow and myocardial pH and lactate concentrations were measured during 60 minutes, and the perfusable proportion of the ventricular wall determined after 15 and 60 minutes of global, substrate-free anoxic perfusion. Myocardial contracture in similarly treated hearts was indicated by the pressure exerted on left intraventricular balloons. Decreased perfusion pressure proportionately decreased both coronary flow and the extent of perfusion, and delayed the onset of contracture. The slower development of contracture was associated with a measured increase in lactate content and fall in pH within the ventricular wall. During anoxia the subendocardial myocardium progressively lost its capacity to be perfused, commencing at the onset of ventricular contracture, and continuing to extend even after maximum contracture pressures had been developed. Increased coronary perfusion pressures accelerated the rate of development of contracture but significantly reduced the extent of the perfusion defect, probably by opposing vascular compression.

Morphological identification of functional capillaries in the myocardium

This study used acrylic resin as an intravascular marker to demonstrate functional myocardial capillaries after fixation by perfusion. Eight rat hearts were excised and allowed to function as isolated organs perfused with oxygenated Krebs-Henseleit buffer (37 degrees C, 10 kPa) for 10 min. Four were fixed by perfusion (4 min) with 2.5% glutaraldehyde at the same temperature and pressure and then immersion fixed (24 hr). The other four hearts were perfused with 0.2% procaine HCl for 30 sec just prior to similar fixation. Polymerizing low viscosity acrylic resin was injected at 10 kPa pressure into the fixed vascular beds and allowed to cure, then transmural blocks of left ventricular myocardium were prepared for scanning electron microscopy. Total initial coronary flow of fixative after procaine treatment was significantly increased, while in untreated hearts the initial fixative flow rate was closely similar to that of oxygenated buffer. The pattern of capillary perfusion was assessed, and the percentage of capillary profiles filled by acrylic resin were calculated. Following procaine treatment, 95.2% of capillaries appeared functional, whereas without procaine arrest, only 62.0% of capillaries allowed the passage of resin. This study indicates that perfusion fixation with glutaraldehyde stabilizes myocardial structure so that the proportion of functional capillary pathways remains closely similar to that in the beating heart and so that such functional capillaries can be identified in morphological preparations by using a low viscosity intraluminal resin marker.

A quantitative study of the sequence of topographical changes in the organ of Corti following acoustic trauma

The organs of Corti of 30 guinea pigs were examined quantitatively by scanning electron microscopy either immediately or 1, 3, 7 or 14 days after exposure to 3 kHz at 125 dB SPL for 30 min. Lesions (0.1-4.15 mm in length) were observed in 70% of the organs of Corti. There was no significant change in lesion length with recovery from the exposure. Early changes in hair cells consisted of stereocilia abnormalities, predominantly amongst inner hair cells and the first row of outer hair cells. The proportion of affected cells increased towards the centre of lesions, where supporting cells were affected also. Subsequent to exposure, affected hair cells were either lost or remained with stereocilia abnormalities but did not recover. Regions showing supporting cell damage were replaced within 3 days by inner sulcus and Claudius cells. Despite similar changes to stereocilia, inner hair cells were more resistant to necrosis than outer hair cells, suggesting that the nature of stereocilia damage does not necessarily indicate the fate of hair cells.

Changes in vascular morphology associated with the no-reflow phenomenon in ischaemic myocardium

To investigate the pathogenesis of the reperfusion defect which develops in ischaemic myocardium, intravascular casts were prepared by injection of methyl methacrylate into the coronary arteries of isolated heparinised rat hearts. Using a scanning electron microscope, the vascular morphology following 60 min of global ischaemia at 37 degrees C was compared to that of non-ischaemic control hearts injected immediately after stopping perfusion with oxygenated Krebs-Henseleit buffer. Complete casts were obtained from control hearts and from all parts of ischaemic hearts except the subendocardial half of the left ventricular wall of ischaemic hearts where the blood vessels were not filled. At the border between the perfused subepicardial and unperfused left subendocardial regions, the resin which filled the radial penetrating arteries and their branches projected from the filled capillary plexus to an extent proportional to their diameter. Intravascular events such as erythrocyte plugging and thrombosis were excluded as causative factors by the use of a cell-free perfusate. Also, there was no morphological evidence that endothelial cell swelling or constriction of any particular population of vessels was involved. The observed pattern of vascular occlusion suggests that, during global ischaemia, blood vessels in the endocardial half of the left ventricular myocardium lose their ability to be reperfused because of extravascular compression.

Histological assessment of orthotopic aortic valve leaflet allografts: its role in selecting graft pre-treatment

Histopathological studies of human cardiac valve grafts recovered at autopsy or re-operation, together with long-term clinical follow-up of valve graft recipients, have indicated that the success of grafts is largely dependent upon the extent to which they are replaced by host fibrous connective tissue. To find the valve preparation technique with least inhibitory effect on tissue ingrowth after grafting, various sterilizing and storage procedures were evaluated using a series of aortic valve leaflet allografts in dogs. To facilitate evaluation, a method for rapidly assaying relative degrees of colonization of grafts was first devised. Application of this method has unequivocally identified a newly-formulated antibiotic solution as the pre-treatment most compatible with host tissue ingrowth.

A comparison of the effects of ischemia and of selective occlusion of capillaries, pre-capillaries and terminal arterioles on coronary reperfusion

To study the 'no-reflow' phenomenon is ischemic myocardium, the effects of ischemia and selective embolic blockade of capillaries, precapillaries and terminal arterioles were compared in isolated rat hearts. Hearts received oxygenated Krebs-Henseleit buffer for 10 min via an aortic cannula, and then coronary perfusion was stopped. The pattern and extent of reperfusion after 15-90 min of global ischemia and after the injection of 9, 15 or 55 mu diameter microspheres were determined from the distribution of injected 6.7% fluorescein isothiocyanate-dextran in frozen transverse sections of the ventricles. Following ischemia, progressively larger subendocardial regions surrounding the left ventricle could not be reperfused. In contrast, embolic occlusion of capillaries, precapillaries or terminal arterioles caused a transmural reduction in perfusion and a fine linear or herringbone pattern of fluorescence. Sixty min of ischemia followed by microsphere injection had no effect on the subendocardial zone of no-reflow but much reduced the intensity of fluorescence elsewhere. Thus thrombosis, erythrocyte plugging and occlusion of capillaries, precapillaries or terminal arterioles are unlikely to be primary causes of the reperfusion defect which develops in ischemic myocardium.

A method for the gross and microscopic investigation of vascular reperfusion in ischemic myocardium

Fluorescein-isothiocyanate dextran (FITC-dextran; MW approximately 70,000) was used in isolated rat hearts to compare normal vascular perfusion of ventricular myocardium with the pattern and extent of reperfusion following 60 minutes of global ischemia. Its gross distribution in frozen transverse sections through the ventricles was similar to that of sodium fluorescein. However, unlike 0.1% sodium fluorescein, 6.7% FITC-dextran has a viscosity similar to that of blood, and its much higher molecular weight prevents its diffusion beyond the ischemically injured vessels. Furthermore, staining by the alcoholic periodic acid-Schiff technique enabled tracer distribution to be confirmed microscopically and distinguished competent from incompetent vessels in paraffin embedded material.

The influence of adenosine on the no-reflow phenomenon in anoxic and ischaemic hearts

Adenosine and the adenosine deaminase inhibitor erythro-9-(2-hydroxy-3-nonyl) adenine (EHNA), separately and in combination, were added to the perfusate of isolated rat hearts which were then subjected to ischaemia or anoxia. The effect of these infusates on the vascular competence of the myocardium subjected to oxygen deprivation of from 15-90 min was compared to control hearts. Vascular competence at selected time intervals was assessed from the distribution of injected. 1% sodium fluorescein in the cut surface of the ventricles. A region of non-perfusion surrounding the left ventricular lumen and involving 25% of the ventricular myocardium developed within 15 min of anoxic perfusion, with little change thereafter. Adenosine had no significant effect on this. The no-reflow phenomenon following ischaemia had similar distribution, but developed more slowly and eventually involved twice as much of the myocardium (59% after 90 min). Surprisingly, pre-treatment with adenosine greatly increased (from 14-46%) the extent of no-reflow after 30 min of ischaemia. Pre-treatment with EHNA caused a slight reduction (59-43%) in its extent but only after 60 min. Thus the no-reflow phenomenon which developed in ischaemic myocardium is unlikely to be due to the reduced vasodilatory action of adenosine.

Histological evaluation of transannular autografts of pericardium placed in the right ventricular outflow tract of dogs

Autografts of pericardium were inserted as patches into incisions across the annulus of the pulmonary valve in 12 adult dogs and 12 pups. In order to accommodate for possible effects of pulmonary hypertension which pertains in many clinical circumstances, constricting teflon bands were placed around the main pulmonary artery in half of the animals. Samples of tissue from the patches were examined by light and scanning electron microscopy from 6 to 14 mth later. The eventual thickness of the insertion site of the patches was found to become 2-14 times greater than the initial autograft due to apposition of layers of collagenous tissue on the endocardial and epicardial surfaces of the autografts. Three pups showed small, angiographically undetected central aneurysms in their patches but the tissue in their base was 5 or more times thicker than control pericardium. Cartilaginous metaplasia was evident at the margins of 57% of the patches, necrotic remnants of parts of the autograft were evident in 38% and foci of dystrophic calcification in 24% of them. Whereas smooth muscle cells sometimes extended into the repair tissue (19%), elastic fibres were not present in the patches. Thus pericardial autografts in the right ventricular outflow tract seem unlikely to develop clinically significant abnormalities.

The effect of an isovolumic left ventricle on the coronary vascular competence during reflow after global ischemia in the rat heart

During global ischemia in isolated rat hearts, the development of contracture, due to irreversible myofilament sliding, causes reduction of left ventricle luminal volume. Also, a considerable area of the myocardium cannot be reperfused after 1 hour's global ischemia. The purpose of this study was to reduce myofilament sliding by placing a fluid-filled isovolumic balloon in the left ventricular cavity of isolated rat hearts and assess the extent of reflow, after 60 minutes' ischemia, by perfusion of a 1% fluorescein tracer solution. Light and electron microscopy was used to determine the state of the vasculature and myofibrillar apparatus. In hearts without the left ventricular balloon (control) the ischemia produced a no-reflow zone comprising 45% of the myocardial wall. In contrast, if an isovolumic balloon was in place during the ischemic period, only 6% of the wall was involved. The volume of the capillary bed in the subendocardium of the control hearts was about 60% of that in th isovolumic hearts. In the isovolumic ("isometric") mode, ischemic contracture was associated with more severe myocardial cell injury than in the corresponding control ("isotonic") mode. Our results support the concept that intramyocardial pressure generated by ischemic contracture plays a major role in the production of the no-reflow phenomenon in globally ischemic rat hearts, and indicate that it is the series elastic component of cardiac muscle which imparts the stiffness necessary to prevent reopening of coronary vessels after a severe ischemic insult.

The effects of altered cation balance on the fine structure of hypoxic myocardial cells

To determine whether the changes in intracellular/extracellular cation balance which develop in ischaemic myocardium are responsible for the fine structural changes seen in such tissue, thin slices of normal canine ventricle were incubated under hypoxic conditions at 37 degrees C and physiological pH in balanced salt solution (BSS), isotonic NaCl and isotonic KCl. Slices from each solution were fixed at 10-120 min intervals and examined by light and electron microscopy. For 60 min, tissue from both NaCl and KCl showed good overall preservation of cell architecture and only mild subcellular alterations including aggregation of nuclear chromatin, disappearance of glycogen granules, and swelling of sarcoplasmic reticulum. Tissue from BSS showed early development of intramitochondrial dense inclusions together with focal contraction-band damage similar to that seen in temporarily ischaemic, re-perfused heart muscle and at the margins of infarcts. These changes thus appear to be promoted by divalent cations. The progressive reversal of monovalent cation balance in an area of permanent and severe ischaemia does not appear to be a major determinant of fine structural change.