Thompson v The Queen

The Crown called evidence from Dr. Verzosa, a forensic medical officer, who said that a hole in Mirjana's skull, which was severely fractured, was in his opinion caused by a .22 calibre bullet fired at very close range to the head. He also stated that the absence of carbon particles and soot deposits in the tracheal/bronchial structure indicated that her death occurred before her incineration. Likewise, Dr. Verzosa thought that either severe head fractures in Ljiljana's skull or a firearm injury caused her death. He did not consider that incineration was the cause of death or the cause of injuries discernible in her skull. There was other expert evidence from Mr. Smith, an analytical chemist called by the Crown, to the effect that the arms of Ljiljana, burned cloth taken from her face, a piece of glass found in her skull and her bracelet all contained residues of a petroleum product. Mr. Smith also found such residues in bits of clothing found on the body of Mirjana.

This evidence acquires greater significance in the light of the evidence of contradictory answers given by the applicant to the police relating to the presence of a petrol container in the back of the car. Initially the applicant told Sergeant Moebus on 5 January 1982 that he was carrying an empty container in the back of the car when it collided with the tree. He conceded then that it had previously been used to carry petrol and he denied that there were any fuel leaks in the car.

Subsequently on 25 October 1984 he told Detective Sergeant Ninness that the container was not empty at the time but contained about five litres of fuel when the car collided with the tree, that he had filled it with fuel for his mother's lawn mower in the afternoon at a garage and that the container was placed behind the driver's seat. The applicant told Detective Sergeant Ninness that petrol spilling from the container was the cause of the interior of the car suddenly catching fire. He also told Ninness that he had smelled petrol just after the collision. He explained his earlier and incorrect assertion to Sergeant Moebus that the container was empty by making the excuse that he did not want to be blamed for "firing the car".

The presence of petrol in the back of the car perhaps explains how witnesses who came upon the scene after the collision initially saw flames coming out from under the bonnet and subsequently saw flames flare up in the passenger compartment of the car, though the burning of the car as related by the witnesses was equally and perhaps more consistent with the car, including the roof, having been doused with petrol. It is significant that the two witnesses who were the first to arrive at the scene noted flames under the bonnet, and smoke rather than flames in the interior of the car. At that time the applicant was not at the scene. He had evidently walked into the bush. This was at a time when it might have been possible to rescue the deceased before the interior of the car was engulfed in flames. The applicant returned to the scene when the flames engulfed the interior of the car.

The Crown also called expert evidence, the effect of which, if accepted, was to establish that the condition of the car was consistent only with its speed being very much less than 40 to 45 miles per hour at the time of the collision. Likewise, expert evidence was called to show that in the light of the quantity of petrol in the engine compartment of the car and the lack of damage to the petrol tank, the applicant's account of how the fire spread was unlikely to be true.

The importance of the similar fact evidence is that, if properly received, it strengthened the Crown case that the applicant shot the deceased. Detective Station Sergeant Lanham was permitted to give evidence that the applicant had shot the four persons in the house at Richardson and the circumstances surrounding the four murders were proved. The trial judge conducted a voir dire hearing. The Crown gave details of the evidence which it intended to lead. His Honour decided to admit the evidence upon the ground that there was a "striking similarity" between the facts of the Richardson murders and the present case and that the "similar facts" were relevant as tending to show otherwise than via mere propensity that the applicant committed the acts forming the basis of the charge because on other occasions he had committed a crime in a particular and rather unusual manner.

Nine points of similarity were suggested by the prosecution and referred to by his Honour. They were:

(1) the family connection between each set of victims;

(2) each of the six victims involved in the four Richardson murders and those for which the accused is being tried, was shot;

(3) each such victim was shot with a similar calibre rifle;

(5) each Richardson victim was shot at close quarters and it was highly likely that the two victims of the present charges were also shot at close quarters;

(6) in respect of the Richardson murders the accused had attempted to destroy the evidence and he had done likewise in relation to the present offences;

(7) the attempt to destroy evidence was in both instances by fire;

(8) in both instances the fire had been set by using petrol (and in relation to the Richardson murders other accelerants); and

(9) the accused was present at the scene of each set of killings when or shortly before they occurred and was the last person to see the victims alive.

We shall consider the sufficiency of these points of similarity and the admissibility of the similar fact evidence later in these reasons.

Despite an application by counsel for the applicant that he should do so, the trial judge did not seek a special verdict from the jury on the issue of jurisdiction. It was common ground that the jurisdiction of the Supreme Court of the Australian Capital Territory to try the applicant depended upon the occurrence within the geographical limits of the Australian Capital Territory of either the deaths of the two girls or the act or acts causing their deaths. Section 25 of the Crimes Act 1900 N.S.W., as amended in its application to the Australian Capital Territory, provides:

Where, in any case of homicide, the cause of death happened on the sea, or elsewhere without the Territory, but the death was within the Territory, or the cause of death happened within the Territory, but the death was on the sea or elsewhere without the Territory, the offence may be dealt with, in all respects, as if the same had been wholly committed within the Territory. The trial judge directed the jury that, amongst other things, they could not convict the applicant unless they were satisfied beyond reasonable doubt that the deaths of the deceased took place in the Australian Capital Territory. The trial judge directed the jury's attention to the various considerations which were relevant to that question.

On appeal the Federal Court rejected the submission that the trial judge should have sought a special verdict from the jury on the question of jurisdiction. Whilst acknowledging that it was permissible to seek such a verdict, the Federal Court concluded that it was not the only proper way to proceed. The Federal Court also rejected the applicant's submission that there was insufficient evidence to enable the jury to satisfy itself beyond reasonable doubt that the deaths occurred in the Australian Capital Territory. Their Honours did not elaborate beyond saying: "[T]here was evidence upon which the jury could find as they must have done that the deaths occurred in the A.C.T. and not in New South Wales."

The applicant has reiterated these submissions in this Court. A special verdict on the issue of jurisdiction was sought in Reg. v Graham [1] but there is nothing in that case which suggests that there is a rule of law requiring a trial judge to seek a special verdict on the issue whether death or the act causing death occurred within the jurisdiction. In that case, as in the present case, it was assumed by the trial judge and the Court of Criminal Appeal that the criminal standard applied to proof that the crime had been committed within the territorial jurisdiction. If that assumption be correct, the trial judge has a discretion to seek a general or special verdict and it has not been shown in this case that his Honour's exercise of discretion in seeking a general verdict was vitiated by any error of law or fact. Different considerations would need to be taken into account if the civil standard were the applicable standard, a question to which we shall return later.

The submission that the evidence could not satisfy the jury beyond reasonable doubt that the deaths occurred within the jurisdiction has rather more force. No bullets, cartridges or firearm were found at the scene of the collision between the car and the tree. Furthermore, a person who intended to kill the deceased might select a more secluded venue rather than run the risk of being observed by travellers on the Monaro Highway, which is the main road between Canberra and Cooma. There are secluded areas in New South Wales adjacent to the Australian Capital Territory accessible to a car travelling to or from Bredbo which would have served the purpose. One such area was the old Cooma Road over the railway line which was fairly close to the scene of the collision. The boundary between the Australian Capital Territory and New South Wales is very close to the railway line and follows it at this point.

On the other hand, there is no evidence which connects the deaths of the deceased with New South Wales. The applicant has never suggested that they died in New South Wales or that he deviated from the Highway on his journey to or from Bredbo. Moreover, a person who planned to kill the deceased would not readily run the risk of killing them away from the scene of the collision as that would entail driving the car for some distance with the bodies of the victims in the car. The absence of a firearm and cartridges is explicable on the hypothesis that they were concealed by the applicant in the bush during his absence from the scene and were picked up by him later.

In these circumstances we for our part have some doubt that the evidence does establish beyond reasonable doubt that the deaths of the deceased occurred in the Australian Capital Territory. On the other hand, we consider that if the civil standard were applicable the preponderance of evidence favours the conclusion that the deaths of the deceased took place in the Territory. Rather than deciding whether or not the jury's verdict on this issue can be supported, it is convenient to consider whether the criminal standard or the civil standard applies to the proof of a crime having taken place within a particular jurisdiction.

There is a dearth of authority in England and Australia dealing with the standard of proof to be applied in proving that a criminal offence has been committed within the territorial jurisdiction of a court. In Graham the question seems not to have been raised for decision by the arguments presented to the court. In Reg. v Hildebrandt [2] Herron C.J., with whom Ferguson and Nagle JJ. agreed, observed:

Where the jurisdiction of the court is challenged on merely territorial grounds the onus is on the Crown to establish beyond reasonable doubt that the crime was committed in New South Wales. It is only in this sense that it is proper to say that proof of the place of the crime is a necessary ingredient of the offence. 1. (1963) 81 W.N. (Pt 1) (N.S.W.) 143, at pp. 150-151.

Hildebrandt had previously been indicted and convicted in Queensland of the offence of putting explosive substances in an aircraft with intent to destroy the aircraft. The Court of Criminal Appeal in that State had set aside the conviction on the ground that the Crown had not proved beyond reasonable doubt that the acts constituting the offence charged were performed in Queensland: see Reg. v Hildebrandt [3] , per Stanley J., with whom Mack J. agreed [4] . Stable J. said simply that the conviction should be set aside due to "a real doubt as to the territorial jurisdiction of the District Court" [5] . More recently, in Reg. v Collins [6] the trial judge gave a direction on the question of jurisdiction based upon the criminal standard and the Court of Criminal Appeal appeared not to question this aspect of the direction. It appears that the issue of what is the applicable standard of proof was not argued in the Supreme Courts of Queensland, New South Wales or South Australia.

1. [1964] Qd R. 43, at p. 50. 2. [1964] Qd R., at p. 54. 3. [1964] Qd R., at p. 63. 4. (1986) 44 S.A.S.R. 214.

The question has been discussed in the United States where the preponderance of authority appears to favour the proposition that it is for the State bringing the prosecution to prove beyond reasonable doubt that the offence occurred within its territorial boundaries: State v Jackson [7] ; State v Baldwin [8] ; State v Batdorf [9] ; Lane v State [10] . Underlying these decisions are two notions: first, that jurisdiction, when contested, is a matter as much to be proved as the prerequisite of the authority of the court to enter judgment (Batdorf [11] ; Jackson [12] ); and, secondly, that the commission of the crime within the jurisdiction is as much an ingredient of the offence as the commission of the crime itself (Jackson [12] ). The commission of the crime within the jurisdiction is said to be an ingredient of the offence because perpetration in violation of the law of the territory is essential to guilt: Jackson [13] .

1. (1917) 77 So. 196. 2. (1973) 67 A.L.R. 3d 979. 3. (1977) 238 S.E. 2d 497. 4. (1980) 388 So. 2d 1022, at p. 1029. 5. (1977) 238 S.E. 2d, at p. 502. 6. (1917) 77 So., at p. 198. 7. (1917) 77 So., at p. 198. 8. (1917) 77 So., at pp. 196-197.

However, this approach has not been adopted universally by courts in the United States. Thus, in People v Cavanaugh [14] , it was held that territorial jurisdiction could be established by a "preponderance" of the evidence. In other cases the courts have regarded the question of jurisdiction in the same way as the related question of venue within the jurisdiction, and have decided the issue according to the lower standard: Cauley v United States [15] ; Younge v United States [16] .

1. (1955) 282 P. 2d 53, at p. 59. 2. (1966) 355 F. 2d 175. 3. (1917) 242 F. 788.

In Canada, the test at least in superior courts appears to be whether the applicant can displace the presumption of regularity, that presumption applying to a superior court exercising criminal jurisdiction: Balcombe v The Queen [17] . In that case Fauteaux J. held that the question of jurisdiction was for the judge and not the jury to decide. In reaching this conclusion, Fauteaux J. [18] drew a distinction between facts relating to guilt or innocence and facts relating to jurisdiction.

1. [1954] S.C.R. 303. 2. [1954] S.C.R., at pp. 305-306.

In the absence of binding authority and compelling precedents from other jurisdictions, the question must be answered in Australia by reference to fundamental principle. Chief Justice Coke's famous definition of murder has been thought by some to imply that the commission of the offence within the jurisdiction is an essential element in the definition of the crime. Coke C.J. said:

Murder is when a man of sound memory, and of the age of discretion, unlawfully killeth within any county of the realm any reasonable creature in rerum natura under the king's peace, with malice fore-thought, either expressed by the party, or implied by law, so as the party wounded, or hurt, &c. die of the wound, or hurt, &c. within a year and a day after the same. [19] But the definition is capable of being read as an assertion that the existence of jurisdiction is a prerequisite to the authority of the court to enter judgment, rather than a discrete element in the definition of the offence itself. Of course, there is no reason to suppose that the standard of proof of jurisdiction is different in the case of murder from the standard of proof of jurisdiction in the case of other criminal offences. So Coke C.J.'s definition does not determine the question. 1. (1797) 3 Inst. 47.

The issue of guilt is necessarily determined within a particular jurisdiction. But the issue cannot be determined unless the prosecution establishes the authority of the jurisdiction to enter judgment. This issue, namely, whether the offence was committed within the jurisdiction, is distinct from that of guilt, namely, whether the elements of the offence are made out. Proof of jurisdiction is a prerequisite of guilt but otherwise it is not an element in proof of the commission of the offence except in those cases in which the offence is so defined that commission of it in a place or locality is made an element of the offence charged. Proof of the commission of the offence must be demonstrated beyond reasonable doubt. But this does not mean that proof of the existence of jurisdiction must first be established beyond reasonable doubt.

The fundamental principle of our criminal law is that the accused's guilt must be established beyond reasonable doubt. The law requires that standard of proof of the commission of a criminal offence in order to eliminate or minimize the chance that an innocent person might be found guilty with all the grave consequences that such an erroneous condemnation would have for the accused, for our system of justice and for the community generally: Brown v The King [20] . The fundamental principle is not offended if the facts essential to the existence of jurisdiction in the court to enter judgment are required to be established according to the civil standard of proof. That is a discrete question which may be left to a jury upon the lesser standard of proof without diverting them from the standard which they are otherwise required to apply in determining guilt or innocence. Cf. Ahern v The Queen [21] . The policy or purpose which underlies the fundamental principle is sufficiently served and the protection of the accused adequately assured if the criminal standard of proof is applied to all the facts relied upon to make out the elements of the offence. To apply that standard to the proof of facts establishing the jurisdiction of the trial court would extend the protection of an accused person to the point of entitling him to an acquittal on the ground that the prosecution could not prove beyond reasonable doubt that the offence was committed in one State or Territory rather than another, even though, if jurisdiction were assumed, the circumstances would be such as to show beyond reasonable doubt that the accused committed the offence charged. To extend the protection in this way would travel beyond the interests which the law seeks to safeguard in imposing the criminal standard of proof and at the same time adversely affect the public interest in the administration of justice by allowing a wrongdoer to escape conviction, notwithstanding that the balance of probabilities suggests that the wrongdoer is subject to the jurisdiction from which he seeks to escape. A wrongdoer clearly subject to the laws of one of two jurisdictions would escape the laws of both, even where such laws were identical, simply because the prosecution could not prove the place of the commission of the offence beyond reasonable doubt. The prospect of this outcome would be lessened if the civil standard of proof were to be applied.

1. (1913) 17 C.L.R. 570, at pp. 584-585. 2. (1988) 165 C.L.R. 87, at pp. 103-105.

Notwithstanding differences in statute law in the various States and Territories (including differences as to penalties), the mutual interest of the States and Territories requires that in a case where one State or Territory is proved to be more likely than not to possess jurisdiction, the courts of the other States and Territories recognize this as sufficient, since it could not be intended that the competing rights of the different jurisdictions combine to prevent any of their laws being enforced. Such an approach accords with the notions of justice expressed by Arterburn J. of the Supreme Court of Indiana, when he observed in State v Carrier [22] :

No person should escape punishment for murder because he is so clever as to conceal the instrument with which the fatal blow was struck or the place where the victim was killed or died. 1. (1956) 134 N.E. 2d 688, at p. 691.

It may be objected that the application of the civil standard of proof to the issue of jurisdiction will serve to confuse the jury and that this is a good reason for applying the criminal standard of proof to that issue. The objection is without substance. There is no reason to think that a jury will fail to understand and act upon a precise instruction which distinguishes between the two standards and makes the point that the civil standard applies only to the issue of jurisdiction, whereas the Crown must satisfy the criminal standard on all other issues.

In some jurisdictions in the United States, courts have applied a rebuttable presumption that the deceased person died within the State or Territory where the body was discovered: see, for example, Commonwealth v Costley [23] ; People v Peete [24] ; United States v Rees [25] ; State v Fabian [26] . But the application of such a presumption does not adequately resolve the problem of jurisdiction.

1. (1875) 118 Mass. 1, at p. 26. 2. (1921) 202 P. 51, at p. 64. 3. (1961) 193 F. Supp. 849, at pp. 859-860. 4. (1972) 263 So. 2d 773, at pp. 775-776.

Although as a matter of logic it will often be possible for a jury to conclude in the absence of evidence to the contrary that a person died in the State or Territory where the body was discovered, this will not always be the case, especially when the body is found close to a border. In other cases the circumstances will be so vague or the facts so contentious that an inference cannot safely be drawn.

There are other objections to the recognition of the presumption, apart from the shortcomings already mentioned. First, because the great majority of homicide cases involve the discovery of a body, the onus of proof of jurisdiction would effectively shift to the accused and tend to undermine the privilege against self-incrimination. The following passage from the judgment of Gray C.J. in Commonwealth v Costley [27] highlights the point:

Evidence having been introduced, strongly tending to show that the homicide was committed by the defendant in Norfolk, the jury were rightly instructed that if they thought that, if it had been committed elsewhere, the defendant would have the means of showing it by other witnesses, they might consider the absence of evidence that it was committed in another county. The evidence referred to related primarily to the state and location of the body. In any event, it would be unrealistic and unfair to expect that an accused person would avail himself of the opportunity to prove that he committed a murder outside the jurisdiction in which he was charged. 1. (1875) 118 Mass., at p. 27.

Secondly, the existence of the presumption would not avoid the need to determine the correct standard of proof, especially in cases such as the present. After application of the presumption, the question remains: what does the accused have to show in order to overcome the presumption?

Finally, although not strictly relevant to the present case, the presumption is of no relevance to the broad range of criminal offences other than homicide, which serves to indicate its failure to address the more fundamental issue of standard of proof.

Accordingly, we conclude that the standard of proof applicable to the establishment of the jurisdiction of a criminal court is the civil standard, that is, upon the balance of probabilities. On this footing we are satisfied that the jurisdiction of the Supreme Court of the Australian Capital Territory to enter judgment in the present case has been demonstrated.

Clearly it will be desirable to seek a special verdict upon the issue of jurisdiction if it is raised. That course was not taken in the present case and a verdict of guilty was returned, but a special verdict indicating lack of jurisdiction will be relevant, if for no other reason, in the event of a further prosecution in a jurisdiction where the defence of autrefois acquit or double jeopardy may be raised. The scope of such a defence was not explored in argument and we refrain from further comment upon it. The special verdict should be sought before any general verdict is taken and if it establishes lack of jurisdiction the jury should thereupon be discharged without taking a general verdict.

We therefore turn to consider the similar fact evidence. The existence of the similar facts which were admitted in evidence by the trial judge was not disputed by the applicant. Indeed, at the trial the applicant's counsel led evidence in the cross-examination of Detective Station Sergeant Lanham that the applicant had been tried for the murder of the four persons killed at Richardson and had been found guilty in October 1984. It was of no consequence that the murders at Richardson were subsequent to the events in question in this case: see Makin v Attorney-General (N.S.W.) [28] . What was disputed was the admissibility in evidence of those facts, it being submitted that they lacked sufficient probative force to justify their admission, having regard to the inevitable prejudice which they raised against the applicant.

The principles upon which similar fact evidence may be admitted have in recent years been examined by this Court in Markby v The Queen [29] ; Perry v The Queen [30] ; Sutton v The Queen [31] ; and Hoch v The Queen [32] . It is established that similar fact evidence ought not be admitted if it tends to show only that the accused has committed another offence or other offences. Proof of the commission of other offences, without more, merely demonstrates a criminal propensity and the prejudicial nature of evidence of this kind is greater than any relevance which it might have. As was observed in Sutton [33] , to admit such evidence would be to invite the jury to proceed upon prejudice or suspicion rather than proof. But if the evidence of the other offence or offences goes beyond showing a mere disposition to commit crime or a particular kind of crime and points in some other way to the commission of the offence in question, then it will be admissible if its probative value for that purpose outweighs or transcends its merely prejudicial effect. The cases in which similar fact evidence may have sufficient additional relevance to make it admissible are not confined, but recognized instances occur where the evidence is relevant to prove intent or to disprove accident or mistake, to prove identity or to disprove innocent association: see Markby [34] ; Sutton [35] ; Harris v Director of Public Prosecutions [36] .

1. (1978) 140 C.L.R. 108. 2. (1982) 150 C.L.R. 580. 3. (1984) 152 C.L.R. 528. 4. (1988) 165 C.L.R. 292. 5. (1984) 152 C.L.R., at pp. 557, 562. 6. (1978) 140 C.L.R., at p. 116. 7. (1984) 152 C.L.R., at pp. 562-563. 8. [1952] A.C. 694.

Although the admissibility of similar fact evidence is restricted by the requirement that it must have sufficient probative force, the use to which it may be put when admitted is no different to that of other circumstantial evidence. That use was explained by Dixon J. in Martin v Osborne [37] :

If an issue is to be proved by circumstantial evidence, facts subsidiary to or connected with the main fact must be established from which the conclusion follows as a rational inference. In the inculpation of an accused person the evidentiary circumstances must bear no other reasonable explanation. This means that, according to the common course of human affairs, the degree of probability that the occurrence of the facts proved would be accompanied by the occurrence of the fact to be proved is so high that the contrary cannot reasonably be supposed. The circumstances which may be taken into account in this process of reasoning include all facts and matters which form constituent parts or ingredients of the transaction itself or explain or make intelligible the course of conduct pursued. 1. (1936) 55 C.L.R. 367, at p. 375.

In the present case the fact in issue which the prosecution sought to establish by tendering the similar fact evidence was that the deaths of the two sisters were deliberately caused by the applicant and did not occur by accident. It may be put that way, for if the deaths were not accidental, it follows inescapably that the applicant murdered the two girls either by shooting both of them in the head or by shooting one and killing the other by the infliction of some other head injury.

The nine points of similarity advanced by the prosecution and apparently accepted by the trial judge went too far. In considering whether the deaths at Richardson displayed the necessary degree of similarity to the deaths of the two sisters, it was impermissible to include as a fact the matter which it was sought to prove, namely, that the deaths did not occur by accident. The points of similarity accepted by the trial judge assumed that the two sisters were shot through the head, as were the victims of the Richardson killings. That was tantamount to assuming the truth of the fact in issue which the similar fact evidence was tendered to prove. As Brennan J. pointed out in Sutton [38] :

It is a canon of logic, rather than of law, that one cannot prove a fact by a chain of reasoning which assumes the truth of that fact. That canon has a particular application in determining the cogency and hence the admissibility of similar fact evidence. When the Crown seeks to tender similar fact evidence as the foundation for inferring a fact to be proved in a trial, it is erroneous to assume the truth of the fact to be proved in determining the cogency of the evidence. That proposition does not preclude reference to direct evidence of the fact to be proved in determining the cogency of similar fact evidence. Similar fact evidence tending to confirm the existence or occurrence of such a fact may be confirmed by direct evidence of the same fact. We would only add to that passage the comment that there does not seem to be any reason why the evidence of the fact to be proved which confirms the similar fact evidence, or adds to its cogency, should be confined to direct evidence. 1. (1984) 152 C.L.R., at p. 552.

The similarity between the killings at Richardson and those of the two sisters was to be considered in the light of the whole of the evidence. That evidence included the damage to the skull of each of the two girls which was said by an expert medical witness to have been caused, or to be consistent with having been caused, by a shot at close range from a .22 calibre rifle. It included evidence suggesting that the two deaths occurred before the car caught fire. There were also the inconsistent stories told by the applicant about the contents of the petrol container in the car and there was evidence which was consistent with the car and the bodies of the deceased having been doused with petrol. In addition there was evidence that the fire started under the bonnet of the car which, when coupled with evidence that the applicant had removed himself from the immediate vicinity, was inconsistent with any attempt by him to extricate either of the two girls from the car. And expert evidence was given that the collision could not have happened in the way in which the applicant said that it did. Finally, there was the family relationship between all of the deceased and the relationship of the applicant to the family.

Thus there was evidence that the two sisters died as a result of head injuries inflicted upon them, in one case, if not the other, by a shot fired at close range from a small calibre firearm. There was evidence that the bodies were subsequently incinerated in a fire deliberately started or accelerated by the use of petrol. There was evidence that the collision between a tree and the car in which the bodies were found was not accidental but staged. Such evidence, albeit disputed by the applicant, served to strengthen the cogency of the similar fact evidence and to assist in determining whether the similar facts were admissible. Once admitted, those facts were available to enable the jury to determine whether or not to accept the prosecution case, including the disputed evidence. Upon that disputed evidence alone, if accepted, and without regard to the similar fact evidence, it would have been open to the jury to conclude that the two sisters did not die accidentally but were murdered by the applicant. But in considering whether they entertained any reasonable doubt about such a conclusion, the jury were entitled to have before them evidence of the murders committed by the applicant at Richardson. The similarity between the circumstances of those murders and the circumstances of the earlier deaths of the two sisters was sufficiently striking to eliminate coincidence as a reasonable hypothesis and to lead to the conclusion that the two sisters did not die accidentally but at the hand of the applicant. The similar fact evidence was, therefore, admissible.

We would grant special leave and dismiss the appeal.

This is an application for special leave to appeal against a judgment of the Full Court of the Federal Court dismissing the applicant's appeal from his conviction by the Supreme Court of the Australian Capital Territory of the murders of Mirjana and Ljiljana Milosevic. I gratefully adopt the recitation by Mason C.J. and Dawson J. of the material facts. The grounds of the applicant's appeal to the Full Court included the following:

2. That His Honour erred in refusing to direct the jury to find the appellant not guilty when there was no or insufficient evidence to establish that the deaths of the alleged victims occurred within the Australian Capital Territory.

4. That His Honour erred in failing to direct the jury to return a special verdict as to whether or not the deaths of the alleged victims or the acts causing such deaths occurred within the Australian Capital Territory.

The Full Court described these grounds as raising the question of the court's jurisdiction, and so, in a sense, they did. But they also raised the question of the territorial ambit of the law of the Australian Capital Territory which defines the crime of murder. The jurisdiction of a court to hear and determine a charge of a criminal offence and the territorial ambit of a law which creates or defines the offence charged are two distinct questions, as Devlin J. recognized in Reg. v Martin [39] : There is a distinction, in my judgment, between what I may call the nature of an offence and the ingredients which have to be present before an offence is committed at all on the one hand, and, on the other hand, the question of what courts are to assume or are to be given jurisdiction when the offence has been committed. See also Reg. v Treacy [40] per Lord Diplock. There is, nevertheless, a close relationship between a jurisdiction to hear and determine a charge of an offence and the "ingredients" (or elements) of the offence. The relationship between jurisdiction and elements will appear more clearly after an examination of the jurisdiction of courts in a common law system to hear and determine charges of offences allegedly committed outside the territorial limits of the State. 1. [1956] 2 Q.B. 272, at p. 285. 2. [1971] A.C. 537, at p. 559.

In England, criminal jurisdiction in Admiralty extended to the hearing and determination of charges of murder committed on British ships on the high seas, for "it is trite and ancient law that the criminal law of England extends to British ships upon the "high seas" - an expression which in the context of Admiralty jurisdiction includes in addition to the open sea all waters below low-water mark where great ships can go": Oteri v The Queen [41] . The juristic rationalization of that doctrine, as Lord Diplock [42] stated it, "is that at common law a British ship fell under the protection of the sovereign; those on board her were within the King's peace and subject to the criminal law by which the King's peace was preserved". Their Lordships rejected the theory that a ship is a "floating island" and thus notionally within English territory. Jurisdiction was founded on the personal subjection of those aboard to the criminal law of England. By statute, the jurisdiction of the Admiral came to be exercised by the ordinary criminal courts. Jurisdiction to hear and determine charges of offences allegedly committed on British aircraft flying in foreign parts was conferred on British courts by s. 62 of the Civil Aviation Act 1949 U.K. but, in the opinion of Devlin J. in Martin [43] , that section did not apply the general body of English criminal law to persons aboard British aircraft. If that view were correct, the jurisdiction under s. 62 was limited to cases where particular criminal laws were otherwise binding extraterritorially on passengers and crew. In Reg. v Naylor [44] , Lord Parker C.J. held s. 62 to be an offence-creating section which he construed as covering "any acts or omissions which would constitute offences if committed in this country unless, and this is, I think, where Devlin J.'s case is in point, they are contrary to some purely domestic legislation". On this view, s. 62 conferred jurisdiction to hear and determine charges of offences which Devlin J. held not to be justiciable under that jurisdiction.

1. (1976) 11 A.L.R. 142, at p. 146; [1976] 1 W.L.R. 1272, at p. 1276. 2. (1976) 11 A.L.R., at p. 147; [1976] 1 W.L.R., at p. 1276. 3. [1956] 2 Q.B., at p. 285. 4. [1962] 2 Q.B. 527, at p. 528.

Apart from offences committed on ships and aircraft, the English courts have long been vested by statute with jurisdiction over British subjects who commit murder on foreign soil. That was effected by a succession of statutes, culminating in s. 9 of the Offences against the Person Act 1861 U.K., which provided:

Where any Murder or Manslaughter shall be committed on Land out of the United Kingdom, whether within the Queen's Dominions or without, and whether the Person killed were a Subject of Her Majesty or not, every Offence committed by any Subject of Her Majesty, in respect of any such Case, whether the same shall amount to the Offence of Murder or of Manslaughter may be dealt with, inquired of, tried, determined, and punished in England And see Reg. v Azzopardi [45] ; Reg. v Page [46] . Statute apart, however, the common law conferred no jurisdiction on English courts to hear and determine charges of murder committed on land outside England and Wales, whether by British subjects or others: R. v Hooper [47] ; R. v Munton [48] . In Australia, neither s. 9 of the Offences against the Person Act nor any adaptation of it was enacted, perhaps because it was thought that Macleod v Attorney-General (N.S.W.) [49] had denied to the legislatures of the States the power to do so. 1. (1843) 2 Mood. 288 [169 E.R. 115]. 2. [1954] 1 Q.B. 170. 3. [1734] W. Kel. 190 [ 25 E.R. 562]. 4. (1793) 1 Esp. 62 [170 E.R. 280]. 5. [1891] A.C. 455.

Under the Canadian Criminal Code, a question of jurisdiction could have arisen under s. 888 of the 1927 Code (re-enacted 1948, c. 39, s. 38):

Nothing in this Act authorizes any court in one province in Canada to try any person for any offence committed entirely in another province. The Canadian Criminal Code, a law enacted by the Canadian Parliament, is in force throughout Canada. Therefore, the situs of an offence as between two Provinces might have affected the jurisdiction of the courts in one Province to hear and determine a charge of an offence allegedly committed in another but, as the offence is committed against a single Canadian law, no question could have arisen as to the territorial ambit of the law. In Balcombe v The Queen [50] , where a person convicted of murder in Ontario sought leave to appeal against his conviction to the Supreme Court of Canada on the ground that the evidence did not show that the murder was committed in Ontario rather than in Quebec, Fauteaux J. [51] rejected the argument that the jury should have been instructed to determine, as a matter related to jurisdiction and not as a matter related to guilt or innocence, whether, upon the view taken by them of the evidence, they were satisfied beyond a reasonable doubt that either the wounds were inflicted or the death occurred within the Province of Ontario. Holding that "the situs of the crime, in so far as it is related to the question of jurisdiction, was a question exclusively for the Court to determine", his Lordship said [52] : [The jury] are concerned only with the guilt or innocence of the prisoner at the bar. Indeed the lawful fulfilment of their duties rests on the assumed existence of the jurisdiction of the Court to try, at the place where it is held, the accused for the crime charged. They are concerned with facts as they may be related to guilt or innocence but not to jurisdiction. There is nothing under the law entitling them, through the whole course of the execution of their duties, to legally make any other pronouncements but those as to which a general or special verdict is authorized by law. If the locality of a crime is a fact on which jurisdiction to hear and determine a charge depends, I would respectfully agree. The arbiter of jurisdiction is the judge, not the jury. There can be no hearing and determination of the charge by the jury unless the court has jurisdiction to do so. Of course, the judge has jurisdiction to inquire into any facts that are necessary to determine the court's jurisdiction to hear and determine the charge but, in this country, I do not think that jurisdiction to hear and determine a charge depends upon the fact - as distinct from the allegation - that the crime occurred within a particular territory. 1. [1954] S.C.R. 303. 2. [1954] S.C.R., at p. 305. 3. [1954] S.C.R., at pp. 305-306.

The term "jurisdiction" in this context is attended with a certain ambiguity. It may refer to the authority of the court to hear and determine the issues arising in a criminal trial, or it may refer to the power of the court to enter a judgment of conviction and to impose a penalty after hearing and determining the issues. Jurisdiction in the latter sense does depend on the jury's verdict when the accused pleads not guilty to the indictment: the power of the court to enter a judgment of conviction and to punish depends upon a guilty verdict. All the facts on which liability to conviction depends are facts to be found by the jury. In other words, if the charge alleges the commission of an offence against the law administered by the court (the law of the forum), the court has jurisdiction to hear and determine the charge, but when an issue is raised as to the locality of the offence the jury may have to decide the issue in order to determine whether the conduct charged falls within the territorial ambit of the law of the forum. Locality then becomes a fact on which liability to conviction depends. In R. v Johnson [53] the court noted the difference between a plea to the jurisdiction of the court to hear and determine a charge and a plea of not guilty on the ground that the acts charged did not fall within the law applicable to the accused.

Although the locality of an element of the offence charged may affect the power to convict and punish, the jurisdiction of the court to hear and determine the charge depends on what the indictment charges. Ordinarily, there would be no occasion for the judge to find, either on the balance of probabilities or beyond reasonable doubt, that the offence charged was in fact committed within the territory of the forum. The judge is concerned to ascertain only whether the indictment charges an offence against the criminal law of the forum. When an indictment charges an offence as having been committed against the criminal law of the forum so that the court has jurisdiction to hear and determine on a plea of not guilty whether the offence was in fact committed in contravention of that law, and when liability to conviction for an offence against that law depends on the locality of an offence, locality is a fact for the jury to determine.

The criminal courts of a State apply only the criminal law of that State: see Glanville Williams, Venue and the Ambit of Criminal Law, Law Quarterly Review, vol. 81 (1965) 276, at p. 277; 395; 518. So, in the present case, if the murders occurred in the Australian Capital Territory the relevant law is ss. 18 and 19 of the Crimes Act 1900 N.S.W. as applied to the Territory (subject to prescribed modifications) by the Seat of Government Acceptance Act 1909 Cth and the Seat of Government (Administration) Act 1910 Cth. Sections 18 and 19 are not part of the laws of New South Wales; they are part of the laws of the Commonwealth: Spratt v Hermes [54] ; Webster v McIntosh [55] . If the murders occurred in New South Wales, the relevant law is s. 18 of the Crimes Act 1900 N.S.W., a State law. Unlike Canada, the general criminal law in Australia consists in the several laws of the respective States and Territories. And unlike the English law applicable to homicides committed abroad by British subjects, there is no legislation which extends the application of the law of homicide to homicide committed in one State or Territory by residents of another. In the present case, the Supreme Court of the Australian Capital Territory exercised its jurisdiction to hear and determine the issues joined on an indictment charging two counts of murder in contravention of the law of the Australian Capital Territory. It did not have jurisdiction to hear and determine charges of murder in contravention of the laws of New South Wales. Nor did the indictments invite the exercise of such a jurisdiction.

1. (1965) 114 C.L.R. 226. 2. (1980) 32 A.L.R. 603, at p. 607.

Do ss. 18 and 19 of the Crimes Act in their application as the law of the Australian Capital Territory purport to impose a penalty for murder committed outside the Territory? This is not a question of jurisdiction but of the territorial ambit of the law - a question to be answered by construing the statute. Under the common law as it was developed in England, the general rule is that the criminal law applies only in respect of acts committed or omissions made within England. In Cox v Army Council [56] , Viscount Simonds said:

apart from those exceptional cases in which specific provision is made in regard to acts committed abroad, the whole body of the criminal law of England deals only with acts committed in England. And [57] : with rare exceptions the whole body of our criminal law is "domestic" in the sense that it is made for the order and good government of this country and is applicable only to acts done on English soil. This rule may be overridden by statute but, in the construction of an offence-creating statute, the presumption is that the legislature did not intend to proscribe acts done outside the territory of the legislature. Thus, in Air-India v Wiggins [58] , Lord Diplock said: in construing Acts of Parliament there is a well-established presumption that, in the absence of clear and specific words to the contrary, an "offence-creating section" of an Act of Parliament (to borrow an expression used by this House in Cox v Army Council) was not intended to make conduct taking place outside the territorial jurisdiction of the Crown an offence triable in an English criminal court. The presumption against a parliamentary intention to make acts done by foreigners abroad offences triable by English criminal courts is even stronger. As Lord Russell of Killowen C.J. said in Jameson [59] : "One other general canon of construction is this, that if any construction otherwise be possible, an Act will not be construed as applying to foreigners in respect to acts done by them outside the dominions of the sovereign power enacting." See also per Lord Scarman [60] . And in Treacy [61] , Lord Reid said: It has been recognised from time immemorial that there is a strong presumption that when Parliament, in an Act applying to England, creates an offence by making certain acts punishable it does not intend this to apply to any act done by anyone in any country other than England. 1. [1963] A.C. 48, at p. 67. 2. [1963] A.C. 48, at p. 69. 3. (1980) 71 Cr. App. R. 213, at p. 217. 4. [1896] 2 Q.B. 425, at p. 430. 5. (1980) 71 Cr. App. R., at pp. 218-219. 6. [1971] A.C., at p. 551.

The construction of an offence-creating statute is affected by international comity which requires the courts of each state to respect the sovereignty of others. Where a statute creates an offence in order to suppress the harmful consequences of acts or omissions - a "result-crime" rather than a "conduct-crime" in the terminology discussed by Lord Diplock in Treacy [62] - comity does not require the courts of the forum to abstain from punishing breaches of the statute when the harmful consequences occur within the territory of the forum even if the act which causes those consequences is done outside: Director of Public Prosecutions v Stonehouse [63] . Lord Edmund-Davies in that case [64] cited Halsbury's Laws of England for the proposition that: "If a person, being outside England, initiates an offence, part of the essential elements of which take effect in England, he is amenable to English jurisdiction." It may be that when the result of an act is itself an element of the offence, the offence may be taken to have been committed both in the place where the act was done and in the place where the result occurred: see Treacy [65] ; Stonehouse [66] . When the language of an offence-defining or an offence-creating statute is quite general, the statute is construed as embracing conduct in apparent contravention of its terms only if an act is done or an omission is made or a result occurs within the domestic territory. Whether it is the locality of the act or of the omission or of the result which brings conduct within the ambit of the statute is a question of construction, but there must be some local element of the offence. In this sense, it is right to say, as Lord Halsbury L.C. said in Macleod v Attorney-General (N.S.W.) [67] : "All crime is local."

1. [1971] A.C., at p. 560. 2. [1978] A.C. 55. 3. [1978] A.C., at p. 83. 4. [1971] A.C., at p. 562. 5. [1978] A.C., at pp. 66, 77. 6. [1891] A.C. 455, at p. 458.

There was authority for the proposition that murder is committed in the place where the death occurred rather than in the place where the act causing death was committed: R. v Coombes [68] ; Ward v The Queen [69] ; but cf. Reg. v Lewis [70] . The scope of a modern law of homicide, however, is usually defined by statute.

1. (1785) 1 Leach 388 [168 E.R. 296]. 2. (1980) 142 C.L.R. 308. 3. (1857) Dears. & Bell 182 [ 169 E.R. 968].

At the time when Sir Samuel Griffith drafted his Code, he noted (in his letter transmitting the draft) that:

In consequence, perhaps, of the insular position of England, the Common Law appears to contain no provision as to the punishment of an offender in a case where several acts or events are collectively necessary to constitute an offence, and where some only of those acts or events occur within the jurisdiction, the rest occurring out of the jurisdiction; such, for instance, as the case of a man who, standing in Queensland territory, shoots a man standing in New South Wales or vice versa, or a man who sends poison from Queensland to be administered to a man in Victoria or vice versa, or a man who by a false pretence made in Queensland obtains property in New South Wales. This problem was addressed by s. 12 of the Queensland Criminal Code, and s. 25 of the Crimes Act addresses the same problem in cases of homicide: Where, in any case of homicide, the cause of death happened on the sea, or elsewhere without the Territory, but the death was within the Territory, or the cause of death happened within the Territory, but the death was on the sea or elsewhere without the Territory, the offence may be dealt with, in all respects, as if the same had been wholly committed within the Territory. This provision prescribes the facts which must exist before it can be said that the law of homicide of the Australian Capital Territory applies to a murder. The ambit of the law extends to a homicide where either the cause of death or the death or both occur within the Territory. There is no statute which otherwise extends the territorial ambit of s. 18.

As ss. 18 and 19 apply only to murders where the cause of death or the death or both occur within the Territory, a jury cannot convict a person as for a contravention of s. 18 unless it is satisfied that the cause of death or the death or both occurred within the Territory. Similarly, a jury in New South Wales (whose laws correspond to those in force in the Territory) cannot convict a person as for a contravention of s. 18 of the Crimes Act of that State unless the jury is satisfied that the cause of death or the death or both occurred within that State. On their true construction, the respective statutes do not embrace murders where neither the act causing death nor the death occurs within the Territory or State, as the case may be. Thus it was that, in Ward v The Queen , a conviction by the Supreme Court of Victoria of one who murdered another on the southern bank of the Murray River was set aside by this Court for want of jurisdiction when it was held that the southern bank lies in New South Wales.

The necessity for the prosecution to prove the location of the element of an offence which brings it within the ambit of the criminal law of the forum is illustrated by Reg. v Hildebrandt [71] . There the accused was charged, inter alia, with putting explosives in an aircraft with intent to destroy it and depositing explosives in an aircraft in circumstances endangering the aircraft or persons in it. These offences were charged under the Queensland Criminal Code and related to events which occurred on a flight between Sydney and Brisbane. On appeal from his convictions before the District Court, the Court of Criminal Appeal quashed the convictions because the evidence could not establish beyond reasonable doubt that the aircraft was flying over Queensland when Hildebrandt "put" or "deposited" the explosives. Subsequently, Hildebrandt was charged and convicted before the District Court in New South Wales on three counts arising from the same incident: placing an explosive in an aircraft with intent to damage and destroy, having possession of a dangerous thing with intent to do grievous bodily harm and having possession of a dangerous thing with intent maliciously to injure the aircraft. In the Court of Criminal Appeal of New South Wales, the convictions were set aside for error in a direction to the jury, but Hildebrandt sought to avoid an order for retrial by relying on his pleas of autrefois acquit: Reg. v Hildebrandt [72] . The argument failed. Herron C.J. [73] noted that "[t]here was clearly jurisdiction in the [Queensland] District Court to try the case, including the question whether the offence took place within the borders of the State of Queensland". But the verdict of acquittal entered by the Queensland Court of Criminal Appeal was in respect of an offence different from the offence subsequently charged in New South Wales. The plea of autrefois acquit depended on the offence charged in New South Wales (which required proof of its commission in that State) being identical with the offence of which he was acquitted in Queensland (which required proof of its commission in Queensland): see per Herron C.J. [74] . Ferguson J. said [75] :

He was acquitted [in Queensland], not because he had been found not guilty of an offence against the New South Wales criminal law, but because he had not been proved to be guilty of an offence against the Queensland criminal law. 1. [1964] Qd R. 43. 2. (1963) 81 W.N. (Pt 1) (N.S.W.) 143. 3. (1963) 81 W.N. (Pt 1) (N.S.W.), at p. 150. 4. (1963) 81 W.N. (Pt 1) (N.S.W.), at pp. 150-151. 5. (1963) 81 W.N. (Pt 1) (N.S.W.), at p. 152.

In the United States the location of the acts constituting an offence has been treated as an issue of jurisdiction, not as an element of the offence, though the weight of American authority requires proof beyond reasonable doubt of the territorial element of jurisdiction once the question is raised: State v Baldwin [76] ; State v Batdorf [77] . Provided jurisdiction is thus established, when the venue of the trial within the State depends on the locality of an element in the offence, that element can be established on the balance of probabilities: see Batdorf [78] . When a question of territorial jurisdiction is raised, the American practice is to submit the issue of locality to the jury and to direct the jury that the State bears the onus of proof beyond reasonable doubt. This practice has the same effect as treating locality as an ordinary element of the offence.

1. (1973) 67 Am. L.R. 3d 979, at pp. 983-984. 2. (1977) 238 S.E. 2d 497, at p. 502. 3. (1977) 238 S.E. 2d, at p. 504.

In each of the common law countries earlier mentioned, where a statute on its true construction so creates or defines an offence that some element of it must occur within the territory, a prosecution fails unless the locality of that element is proved. Locality is an element of the offence or, if one likes, a fact on which the power to convict depends. It matters not what terminology is used. What is critical is that the accused is not liable to conviction under the statute unless the prosecution discharges the onus of proving locality.

The courts have been conscious of the anomalous lacuna which could occur if the locality of the element which brings an offence within the ambit of a law of the forum had to be proved to the jury's satisfaction beyond reasonable doubt. In State v Carrier [79] , Arterburn J. noted the anomaly:

No person should escape punishment for murder because he is so clever as to conceal the instrument with which the fatal blow was struck or the place where the victim was killed or died. The lacuna would have been largely obviated if the theory propounded by Devlin J. in Martin [80] had been adopted. His Lordship said [81] that crimes conceived by the common law - which are mostly offences against the moral law, such as murder and theft - are not thought of as having territorial limits. They are universal offences. Murder is a crime whether done in France or in England; but if done in France the English courts would not under the common law assume jurisdiction to punish it because that would be an infringement of French sovereignty. The theory is contrary to the limitations which English courts had placed on their jurisdiction to entertain charges of murder committed outside England and Wales. Moreover, the theory that there is no territorial limit on offences which are morally objectionable was rejected by the House of Lords in Board of Trade v Owen [82] . There the prosecution argued, citing Martin , that a conspiracy entered into in England to do out of England acts which are mala in se is an offence, but the argument was rejected: see per Lord Tucker [83] . In this country, a theory that murder is a crime without territorial limits cannot be reconciled with the order made by this Court in Ward v The Queen [84] . 1. (1956) 134 N.E. 2d 688, at p. 691. 2. [1956] 2 Q.B. 272. 3. [1956] 2 Q.B., at p. 286. 4. [1957] A.C. 602. 5. [1957] A.C., at p. 633. 6. (1980) 142 C.L.R. 308.

It follows that, unless the prosecution in a case of homicide proves that the act causing death or the death or both occurred within the Australian Capital Territory, the accused is not liable to conviction as for an offence against s. 18 of the Crimes Act in force in the Territory. Whether the locality of the act causing death or the locality of the death is classified as an element of the offence or as a fact on which the jurisdiction to convict depends, the onus of its proof rests on the prosecution. There remains the question whether that onus must be discharged by proof beyond reasonable doubt.

In argument before this Court, counsel for the respondent conceded that if there were no evidence capable of satisfying the jury beyond reasonable doubt that the deaths of Mirjana and Ljiljana Milosevic or the cause of their deaths occurred in the Australian Capital Territory, then the conviction could not stand. The concession was, I think, too generous.

The golden thread which is always to be seen through the web of the English criminal law (Woolmington v Director of Public Prosecutions [85] ) has been woven through cases in which the locality of an element of the offence was not in issue. Perhaps in England geography eliminated the need for its appearance on issues of locality. In Canada, the standard of proof of locality was not a question likely to arise under a Code which spanned all Provinces. But, as in the United States, the question has arisen and is likely frequently to arise in this country with its several independent law areas, across the borders of which there is a constitutionally assured freedom of movement. Does the anomaly noted by Arterburn J. in Carrier have full rein in criminal prosecutions in Australia? The answer depends upon the reason why proof beyond reasonable doubt is required in criminal cases. In principle, there seems to be no reason why the jealous protection of liberty from which the golden thread is spun should have any part to play when the problem in a particular case is which of two legal systems should punish an offender whose guilt of an offence indifferently punishable by either system is established beyond reasonable doubt.

The conduct which is punishable as an offence under the general criminal laws of the several States and Territories does not greatly vary. Conduct which contravenes the law of one State or Territory will generally be proscribed by the law of other States and Territories. Under the majority of general criminal laws, the locality of an element of an offence is important in determining which State or Territory has authority to punish the proscribed conduct rather than as prescribing the kind of conduct which exposes the offender to liability to be punished. It is one thing to require proof beyond reasonable doubt of every element of conduct which exposes a person to criminal punishment but, when the conduct charged exposes the offender to punishment of the same order whether it has been committed on one side or another of a border, there is no reason why the liability to punishment by one State or Territory rather than another should be proved to the same standard. I would therefore hold that the standard of proof required to satisfy the prosecution's onus of proving locality of an element (when that issue is raised) is generally proof on the balance of probabilities. There is an exception. If the conduct charged is an offence on one side of the border but not on the other or, in one forum, exposes the offender to punishment of a higher order than in another, locality is a fact on which liability to punishment depends. In such a case, the Woolmington rule must apply in all its protective rigour.

There is a further practice matter which should be mentioned in relation to the element of locality, namely, the desirability of seeking a special verdict on that issue when it is raised. Although a jury cannot be required to return a special verdict (Mayor and Burgesses of Devizes v Clark [86] ; Chitty, Criminal Law, vol. 1, 2nd ed. (1826), p. 642), it is desirable for a judge to invite the jury to do so if it should acquit in a case where the locality of an element of the offence is in issue. It is an exceptional case where liability to conviction depends on that issue but, where it does, the jury is competent to return a special verdict on that issue: see Kingswell v The Queen [87] . The desirability of a special verdict as to locality depends on the general verdict which the jury proposes to return. If the jury proposes to convict, the verdict of guilty imports, ex hypothesi, that the jury has found the issue of locality and all other issues joined between the prosecution and the accused against the accused. But if the jury proposes to acquit, a general verdict would not disclose whether the prosecution had failed to prove locality or had failed to prove one of the other issues on which liability to conviction depends. The verdict would be opaque and would not show whether a prosecution founded on the same conduct in a neighbouring jurisdiction could be met by a plea of autrefois acquit. In directing a jury in a case where locality is in issue, a judge should invite the jury to return a special verdict on locality only if it proposes to acquit.

1. (1835) 3 Ad. & E. 506, at pp. 511, 512-513 [111 E.R. 506, at p. 508]. 2. (1985) 159 C.L.R. 264, at p. 287.

In this case there was sufficient evidence, in my view, to satisfy the jury that the murders occurred in the Australian Capital Territory whichever standard of proof be applied. Counsel for the respondent sought assistance from an inference which has been drawn in some of the American cases that a homicide occurs in the locality where the body is found (see People v Peete [88] ; Batdorf [1] ), and submitted that there was evidence in this case sufficient to discharge the Crown's onus of proving the locality of the act of killing or the deaths, bearing in mind that the weight of the evidence adduced by the Crown on that issue had to be assessed "according to the proof which it was in the power of one side to have produced, and in the power of the other to have contradicted": Blatch v Archer [2] . There is force in these submissions. The accused said that they were travelling north from Bredbo along a road which had in fact entered the Australian Capital Territory some 10 to 12 kms from a place where the car containing the bodies was found against a roadside tree. Although the main road ran close to the border for that distance and although it was joined by the old Cooma Road, which crossed the border and led to a secluded spot in New South Wales, there was nothing to suggest that the accused had taken his victims away from the main road. In my opinion, it was open to the jury to find that the accused had killed his victims on or next to the main road at or close to the position where their bodies were found. The trial judge directed the jury to acquit unless they were satisfied beyond reasonable doubt that Mirjana and Ljiljana Milosevic died within the Australian Capital Territory, and the jury's verdict testifies to their satisfaction. I respectfully agree with the Full Court of the Federal Court that there was sufficient evidence to support the convictions.

1. (1921) 202 P. 51, at p. 64. 2. (1977) 238 S.E., at p. 503. 3. (1774) 1 Cowp. 63, at p. 65 [98 E.R. 969, at p. 970].

I agree with Mason C.J. and Dawson J. that the evidence of the subsequent murder of other members of the Milosevic family at Richardson was admissible. All members of the family met violent deaths in the presence of the accused, the violence in every case being consistent with murder by shooting the victim in the head at close quarters with a .22 calibre bullet. The accused had a .22 rifle. In the present case and in the case of the Richardson murders, an attempt was made to destroy evidence of the events which had occurred by a petrol-fuelled fire. There was no evidence of a motive for the Richardson murders which dissociated them from the deaths of Mirjana and Ljiljana Milosevic. In these circumstances, evidence that the accused murdered four members of the family at Richardson by shooting them in the head at close quarters with a .22 calibre rifle and that he tried to destroy evidence of his offence by a petrol-fuelled fire tends to establish that the violent deaths of Mirjana and Ljiljana Milosevic in the presence of the accused and the fire in the car which contained their bodies were events which were caused intentionally rather than by accident. The possibility that the deaths of Mirjana and Ljiljana Milosevic were accidental, though attended by circumstances strikingly similar to the circumstances attendant on the deaths of the other members of the family whom the accused murdered, is unlikely. It would be "an affront to common sense" (to adopt the words of Lord Cross of Chelsea in Reg. v Boardman [3] ) to exclude evidence of the murders at Richardson from the body of proof which tended to show that murder rather than accident was the cause of the deaths of Mirjana and Ljiljana Milosevic.

I would therefore grant special leave to appeal and dismiss the appeal.

The background facts and the issues involved in this application for special leave to appeal are set out in the joint judgment of Mason C.J. and Dawson J. On the question whether the evidence that the applicant subsequently killed other members of the Milosevic family was properly received as "similar fact evidence", I agree with the judgment of Gaudron J. For the purpose of deciding whether similar fact evidence is admissible to prove a disputed fact or circumstance, it is not permissible to assume the existence of that fact or circumstance. In the present case, the issue to which the similar fact evidence was directed was whether the deaths of the alleged murder victims were accidental. Like Gaudron J., I consider that the "similar fact evidence" was admissible not by reason of striking similarity between the circumstances of the subsequent murders and the circumstances of the presence of the bodies of the alleged victims in the burnt-out car which had run into a tree, but by reason of the underlying unity of the two overall incidents in the context of the applicant's admitted involvement in each of them.

On the other question involved in the appeal, I agree with Brennan J. that it was open to the jury to find that the evidence established beyond reasonable doubt either the cause of the deaths of the victims or their actual deaths occurred within the Australian Capital Territory. I also agree with the conclusion of Brennan J. that, in view of the fact that there was no relevant difference between the criminal law of the Australian Capital Territory and the criminal law of New South Wales which would have been applicable if the cause of death and death had occurred in that State, proof of locality was, in any event, only required on the balance of probabilities. However, the reasons which lead me to that last-mentioned conclusion diverge from those of Brennan J. I shall, in explaining my reasons for it, seek to avoid unnecessary repetition of either the effect of the decided cases or the content of relevant statutory provisions.

There are special circumstances in which acts committed outside a particular territory are so connected with it that their proscription and punishment can legitimately be seen as part of both the government of that territory and the government of the territory in which the acts occurred. Putting such cases to one side, it is a general thesis of the common law and an incident of the doctrine of sovereignty under international law that crime is essentially local or domestic: see e.g., Reg. v Keyn [4] . Quite apart from questions of legislative competence, it "has been recognised from time immemorial that there is a strong presumption that" a statute which creates an offence by making certain acts punishable is not intended "to apply to any act done by anyone" in a place outside the territory of the particular national legislature: see per Lord Reid, Reg. v Treacy [5] ; Cox v Army Council [6] ; Air-India v Wiggins [7] . That strong presumption applies, by analogy, to the legislation of Australian States and Territories creating criminal offences. Indeed, it is reinforced in the case of the local law of an Australian State or Territory in that the actual validity of the law will ordinarily depend upon nexus with the particular State or Territory. In the context of that strong presumption, it is clear that the provisions of the Seat of Government Acceptance Act 1909 Cth, in so far as they apply and continue the provisions of the Crimes Act 1900 N.S.W. as part of the law for the government of the Australian Capital Territory, operate to create the crime of murder under the law of the Australian Capital Territory only in circumstances where the cause of death or the death occurred within that Territory: see Crimes Act, s. 25. It follows that the locality of the acts causing death or the death is itself an ingredient or element of the crime of murder under the law of the Australian Capital Territory. If the acts causing death and death all occurred outside the Territory, there has been no offence against that law.

1. (1876) 2 Ex. D. 63, at pp. 68, 117, 152, 160-161, 239. 2. [1971] A.C. 537, at p. 551; at pp. 542, 552, 564. 3. [1963] A.C. 48, at p. 67. 4. (1980) 71 Cr. App. R. 213, at p. 217.

On the other hand, it does not follow that the jurisdiction of the Supreme Court of the Australian Capital Territory to try a charge of murder only exists if it be actually proved that the cause of death or death occurred within the Territory. The criminal jurisdiction of the Supreme Court of the Australian Capital Territory extends to the trial of alleged offences against the law of the Territory. It exists in a case where the accused is charged with murder under that law regardless of whether, in the event, the evidence establishes all or any of the ingredients of the offence charged. Indeed, where the charge is one of murder against the law of the Territory, the jurisdiction of the Supreme Court of the Australian Capital Territory is exclusive. The allegation that the cause of death or death occurred within the Territory which is implicit in a charge of murder against the law of the Territory necessarily suffices to found the jurisdiction of the Supreme Court to deal with the charge. Where locality is an ingredient of the offence charged, the matter is not, however, merely one of jurisdiction. Thus, the evidence will not suffice to prove guilt of the charge of murder under the law of the Australian Capital Territory unless it establishes the existence of all ingredients of the offence, including the locality of the cause of death or the death. That being so, the fundamental requirement of the common law that the ingredients of a criminal offence must be proved beyond reasonable doubt is prima facie applicable to the proof of locality. If the various States and Territories constituting the Commonwealth were, for purposes of identifying the content of internal laws, to be treated as independent nations with their own discrete systems of common law, the conclusion that proof of locality was required beyond reasonable doubt would, in my view, be irresistible since there would be no compelling consideration operating within the confines of the law of the particular State or Territory which would require the punishment of a person for an alleged offence against that law in circumstances where it was not established beyond reasonable doubt that the alleged offender was in fact guilty of any such offence.

In truth, however, the local laws of the various States and Territories of the Commonwealth cannot properly be treated as if they were the discrete systems of law of independent nations. They are the components of a single national legal system: see Breavington v Godleman [8] . and see, also, Sir Owen Dixon, Sources of Legal Authority, Jesting Pilate (1965), pp. 198-200. The Constitution itself embodies that system of law. It assumes the substratum of the common law upon which it was founded. It empowers the Commonwealth Parliament to make laws with respect to the subject matters which it enumerates, including laws for the government of the Territories. It continues the constitutions and laws of the States to the extent that they are consistent with both the Constitution itself and valid laws of the Commonwealth made pursuant to its terms. When inconsistency would otherwise exist between the statutory laws of different elements of the Federation, the Constitution itself resolves it: in the case of inconsistency between a Commonwealth law and a State law, by the paramountcy of Commonwealth law under s. 109; in the case of inconsistency between the laws of different States, by the confinement of the operation of State laws by reference to territorial (or predominant territorial) nexus under the constitutional structure and the mandatory full faith and credit directive of s. 118. Subject to the Constitution itself and to valid statutory provisions, the substantive law of Australia is the common law which transcends internal State or Territorial boundaries and operates as "an entire system" (cf. Sir Owen Dixon, "The Common Law as an Ultimate Constitutional Foundation", op. cit., p. 204). So it is that, within the body politic created by the Constitution, one set of facts will fall to "be adjudged by only one body of law and thus give rise to only one legal consequence, regardless of where in the Commonwealth the matter [falls] for adjudication" (per Wilson and Gaudron JJ., Breavington [9] ). That one body of law is the law of the Australian nation which speaks with a single voice and not as a babel of nine different Commonwealth, State or Territory voices all speaking at the same time but saying different things.

That is not, of course, to suggest that the content of the national law applicable to particular circumstances will be identical regardless of where those circumstances occur within the Commonwealth. Nor is it to deny the importance of the Constitution's qualified preservation of the constitutions, powers, laws, and territorial limits of the federating Colonies. It is simply to acknowledge that the law of a State or Territory itself operates within the national legal system and as part of the national law. Where the Constitution itself or a valid general Commonwealth statute applies directly, the national identity of the system of law is at its most obvious. Where the precise content of the applicable law depends upon territorial or other nexus within the Federation, the national identity of that applicable law may be less manifest. Analysis discloses, however, that the rules for identifying the law applicable to particular circumstances are themselves part of the national law and that the applicable law so identified, regardless of whether it be a State constitution or a law made by a State or Territory legislature, derives its ultimate authority from the Commonwealth Constitution itself and applies as part of the national system of law which that Constitution founded.

What then is the significance for present purposes of the fact that the laws of the various States and Territories of the Commonwealth are but part of an overall national system of law? Of itself, that fact provides no reason why the ordinary criminal onus should not apply to the proof of locality as an ingredient of an alleged offence against a law of a State or Territory to the extent that there is an issue about whether the alleged offence was committed within or beyond national territory. In such a case, the rationale of the requirement of proof beyond reasonable doubt (i.e. the protection of the innocent from wrongful conviction) exists since (special cases aside) the alleged offender will be innocent of any crime against any part of the national law if the acts constituting the alleged offence all took place outside Australian territory. The same can be said of the case where the question is whether the alleged acts took place in the State or Territory in which the accused stands charged or in some other part of the Commonwealth where they would not or might not constitute an offence. Again, in such a case, the ordinary rule that all the ingredients of a crime must be established beyond reasonable doubt should apply since there would not or might not be any crime at all if the alleged acts occurred in that other part of the Commonwealth. The position is, however, quite different in a case where the question involved in an issue about location is whether the alleged acts took place in one or other of a number of different local areas in circumstances where there is correspondence between the relevant criminal law of each in the sense that the alleged acts would constitute a criminal offence under the applicable local law if committed in any of them. In such a case, the alleged acts constitute a criminal offence under the national law regardless of in which of the possible areas of the Commonwealth they were committed and the essential relevance of their location is for the identification of the particular legal provisions under which the alleged offender is liable to be tried and punished.

The present case falls within the last-mentioned category. The issue about location involved the question whether the cause of death and deaths occurred in the Australian Capital Territory or in New South Wales. In all relevant respects, the criminal law of those two parts of the Commonwealth corresponded. If the acts alleged against the applicant had happened in New South Wales, the applicant was guilty of the crime of murder under the law of that State. There is no relevant defence which would have been available to him under the law of New South Wales which was not available under the law of the Australian Capital Territory. The nominal penalty for murder (where the accused is over 18 years of age) is imprisonment for life in both parts of the Commonwealth. There is no applicable statutory provision defining the standard of proof of the locality of the acts causing death or the death in such a case. That being so, the question whether proof beyond reasonable doubt was required falls to be determined by reference to the common law, that is to say, the common law of Australia. There is no judgment of this Court which decides that question. I have found little real assistance in answering it, in an Australian context, in the judgments of other courts.

The requirement of proof beyond reasonable doubt of the ingredients of an alleged crime constitutes an entrenched tenet of the common law of this country. As has been seen, departure from that requirement in so far as proof of locality is concerned cannot be rationalized in the present case on the ground that the locality of the applicant's crimes was merely relevant to the jurisdiction of the Supreme Court. If the acts causing the deaths and the deaths in the present case all occurred in New South Wales, the applicable local component of the national law was the criminal law of that State. The applicant was, in that event, guilty of crimes against the law of New South Wales. He was not guilty of the crimes of murder under, for example, the law of Western Australia. Nor was he guilty of the crimes of murder under the law of the Australian Capital Territory with which he was charged before the Supreme Court of that Territory and of which he was convicted by that court. It follows that a conclusion that proof beyond reasonable doubt of locality was not required in the present case would reflect an essentially pragmatic qualification of general principle. Such a qualification of so fundamental a tenet of the common law could only be justified by compelling considerations of policy and good sense. In my view, such compelling considerations exist.

The rationale of the requirement of proof beyond reasonable doubt is (as has been mentioned) the protection of the innocent from wrongful conviction. To require such proof of the specific location within national territory of a criminal act in circumstances where criminality and liability to punishment correspond under the components of the national law applicable to the competing possible locations would be to derange that rationale and to mock the administration of criminal justice. The result of such a requirement in those circumstances would be the grotesque situation that what was plainly a criminal offence under the national law regardless of the side of some inter-State or State-Territory border on which it occurred could not be punished unless it could be proved beyond reasonable doubt that it was committed on one or other side of that border. In other words, the common law of this country would effectively confer immunity from the reach of the criminal law upon a person who had undoubtedly committed a crime under whatever was the applicable local component of the national law but whose crime was, deliberately or fortuitously, committed in circumstances where there was some reasonable doubt about whether it had been committed on one side or another of a State or Territory border. Considerations of justice, policy and common sense all combine to compel the conclusion that the general requirement of proof beyond reasonable doubt should not apply with respect to proof of locality in such circumstances. Were it otherwise, the result would be that a calculating criminal could effectively convert the vast unmarked internal border areas of this country into modern alsatias for premeditated crime.

It should be acknowledged that acceptance of proof of locality upon the balance of probabilities does not resolve all problems which might arise in a case where the criminal acts of an accused occurred within Australia and the only question involved in an issue of location is whether they occurred in a Territory or a State whose laws relevantly correspond. For one thing, there is a problem about what degree of correspondence should be required between the relevant laws of the competing locations. There is obviously something to be said for the view that proof beyond reasonable doubt should be required in any case where the maximum penalty upon conviction is greater in the location of the trial than under the law of another possible location. For another thing, even a requirement of proof of locality on the balance of probabilities leaves open the possibility that a person who is obviously guilty of a serious crime against the national law will be immune from all punishment because of a fortuitous or contrived absence of evidence about the precise location of his crime. There is something to be said in favour of the recognition of rebuttable presumptions about location, such as that death occurred in the location where the body was first found or known to have been. Those are, however, problems which must be left to another day. In the meantime, it would seem appropriate to draw attention to the desirability of joint Commonwealth and State legislative action to deal with the difficulties involved in this area of the common law of this country.

In the result, the direction of the learned trial judge to the jury to the effect that it was necessary that they be satisfied beyond reasonable doubt that either the cause of the deaths or the deaths occurred within the Australian Capital Territory was unduly favourable to the applicant. The conclusion that the implied allegation of locality was sufficient to found the jurisdiction of the Supreme Court to deal with the alleged offences effectively disposes of any argument that the learned trial judge was obliged to seek a special verdict from the jury on the question of locality.

I would grant special leave to appeal and dismiss the appeal.

I agree with Mason C.J. and Dawson J., for the reasons that their Honours give, that jurisdiction need only be established on the balance of probabilities and not beyond reasonable doubt. I wish to add only the observation that proof of jurisdiction is proof of the occurrence of some act or omission within the jurisdiction, that act or omission being selected by the law as the condition of its operation upon the acts or omissions said to constitute the offence charged. Locality, in this sense, may be contrasted with locality as an element of the offence charged, as, for example, in the case of an offence which is constituted by acts or omissions in a public place. In the latter case, locality is an essential element of the offence and as such may be decisive of criminality; in the former case, locality is decisive only of the operation of the law and of the jurisdiction of the courts charged with administering that law to enter judgment. The issue of jurisdiction is thus a discrete issue and distinct from the issue of guilt which depends upon the elements of the offence charged being made out.

I agree with Mason C.J. and Dawson J. that the evidence in the present case was sufficient to establish, on the balance of probabilities, the facts necessary to ground jurisdiction in the Supreme Court of the Australian Capital Territory. I agree also with their Honours' observations on the taking of a special verdict.

I agree also with Mason C.J. and Dawson J. that evidence as to the applicant's involvement in and the manner and circumstances of the murder of other members of the Milosevic family was properly admitted at the trial of the applicant on charges of the murder of Mirjana and Ljiljana Milosevic. The evidence has been described as "similar fact" evidence, an expression which is descriptive of certain evidence which is admissible by reason of its probative value in "admeasuring the probability or improbability of the fact or event in issue, if given the fact or facts sought to be adduced in evidence": Martin v Osborne [10] ; Hoch v The Queen [11] .

1. (1936) 55 C.L.R. 367, at p. 385. 2. (1988) 165 C.L.R. 292, at pp. 294-295.

The probative force, the criterion of the admissibility of what is referred to as "similar fact" evidence, lies in the evidence revealing "striking similarities", "unusual features", "underlying unity", "system" or "pattern": see Hoch [11] . It is sometimes convenient, but not essential, to identify points of similarity as a step in the evaluation of the probative force of the evidence sought to be adduced. In the present case the probative value lies not so much in "striking similarity", "system" or "pattern", considerations which might necessitate identification of points of similarity, but in "underlying unity".

The subsequent murder by the applicant of other members of the Milosevic family in circumstances involving the attempted incineration of their bodies, as a matter of ordinary experience of human behaviour, was capable of supporting an inference, at the level required for circumstantial evidence, that those deaths were inextricably connected or united with the deaths of Mirjana and Ljiljana, their deaths also involving the applicant and the incineration of their bodies. That inextricable connexion or underlying unity, rather than points of similarity, raises the improbability of the deaths having occurred in a manner other than as alleged in the prosecution case and gives the evidence that probative force which renders it admissible.

I would grant special leave to appeal and dismiss the appeal.