[1950] EACA 453

# H. M. COURT OF APPEAL FOR EASTERN AFRICA

Before SIR NEWNHAM WORLEY (President), BACON, Justice of Appeal and MAHON, Ag. C. J. (Tanganyika)

### CANISIO s/o WALWA, Appellant (Original Accused)

# REGINAM, Respondent

### Criminal Appeal No. 364 of 1955

(Appeal from the decision of H. M. High Court of Tanganyika, Harbord, J.)

# Evidence—Accomplice by incorrect inference—Appellate Court's power to reverse inference—Indian Evidence Act, 1872, section 133.

This was a second appeal from a decision of the High Court of Tanganyika dismissing an appeal from a decision of the District Court at Mwanza. The appeal was first on the ground that the learned Judge's decision constituted an error in law in that he reversed a finding of fact of the trial Magistrate, and secondly that it was not open in law to the Magistrate to convict upon the uncorroborated testimony of persons who were found by him to be and who should accordingly be regarded by both appellate Courts as accomplices.

Held (25-1-56).—(1) That the learned Judge had not reversed a finding of the trial Court which depended to any extent on credibility but had merely substituted the right inference for the wrong one which the Magistrate had drawn which it is open to an appellant Court

to do.

(2) That the second ground of appeal did not arise but the Court gave its view of the general principles of the rules relating to accomplices and corroboration of their evidence.

Appeal dismissed.

Cases referred to: Fazleabbas Sulemanji v. Reg., 22 E. A. C. A. 395; Davies v. Director<br>of Public Prosecutions, (1954) A. C. 378 H. L.; R. v. Haji Mohamed Saleh Mohamed,<br>(1933) 15 L. R. Kenya 109; Kichingeri v. R., (1908) 3 E. A. R. v. Wamjerwa, (1944) 11 E. A. C. A. 93; R. v. Ndaria and others, (1945) 12 E. A. C. A. 84.

### Dodd for appellant.

#### Summerfield for respondent.

JUDGMENT (prepared by Bacon, J. A.).—This is a second appeal from a decision of the High Court of Tanganyika dismissing an appeal against a conviction of stealing certain cotton fabrics at Fela in Lake Province recorded in the District Court of Mwanza on 18th May, 1955.

The appellant was at the material time the station master at Fela. The fabrics in question were packed in three bales and despatched by rail from Dar es Salaam in a goods wagon which was to travel as far as Fela, the bales being destined for three consignees respectively at three other stations on the line. None of the three bales was offloaded at its proper destination and none of the consignees received his consignment. Within a few days, the wagon having by then been detached from the train at Fela and sidetracked to the goods shed at the station, the appellant offloaded the bales from the wagon and delivered their contents to one Gulamhussein who disposed of them for the appellant's account. The taking of the bales and the delivery to Gulamhussein

was effected in two stages, each during the evening, the second within two or three days of the first. On each occasion the appellant personally supervised the operation, assisted by two pointsmen employed at Fela station who, acting under the express instructions of the appellant, did the manual labour of moving and opening the bales and delivering the fabrics.

The case for the Crown thus rested mainly on the evidence of Gulamhussein, of one Ambalal Patel who on Gulamhussein's instructions transported the first bale from the goods shed to a shop in Mwanza, and of the two pointsmen who helped to take and deliver all three bales. The learned Magistrate found that undoubtedly those two first-named witnesses were accomplices. As regards the pointsmen, however, his finding on this question was much less clear, though he did eventually find that they also were accomplices. Being thus faced with the question of looking for corroboration of an accomplice's evidence and its application to the present case, the learned Magistrate clearly expressed the warning as to the danger of convicting without corroboration, held that there was no corroboration here, held that nevertheless he was entitled to convict if he believed the accomplices, and, being "abundantly satisfied that their (the pointsmen's) evidence was truthful", accordingly convicted.

The first appeal was upon the grounds that the learned Magistrate ought not to have believed the pointsmen and that in any event it was an error in law to convict on the uncorroborated evidence of accomplices. In his judgment, however, the learned Judge dealt only with another point which, in view of his decision thereon, alone sufficed to dispose of the first appeal. Having cited in toto the learned Magistrate's findings of fact as to the physical part played by each pointsman and the passages in the Magistrate's judgment discussing the vital question as to whether either pointsman had a guilty mind or was merely acting under the appellant's orders without appreciating the significance of what was being done with the bales, the learned Judge concluded that, in the light of the Magistrate's own observations on the question of mens rea, the latter erred in finding that the pointsmen must be regarded as accomplices at all. In a word, the learned Judge held that the chain of reasoning was broken and that the Magistrate's conclusion was a non sequitur by reason of the following passage in his judgment: —

"At first I was inclined to the opinion that they were mere passive instruments and in no way participes criminis. After further consideration, however, I conclude that it would be more proper to treat them as accomplices. What influences me most in this opinion is the reflection that, had these men been charged with stealing the goods, it would surely have been of no avail to them to plead that they were merely carrying out the instructions of a superior official. I was nevertheless somewhat reluctant to treat them as accomplices for I received a very strong impression that both were speaking the truth and they did in fact no more than obey orders. It is true that they are not 'raw natives' in the words of the E. A. C. A. in Abdulrasul s/o Jivraj, 3 Tang. L. R. 71 at p. 72, but I have had some experience of the almost incredible gullibility and complaisance of the African in the face of real or purported authority. I have, however, decided that I had better regard them as accomplices, for the rule exists for the protection of accused persons and should not be reluctantly invoked....

I received a very strong impression that both were telling me the truth and that they were simply doing what they were told to do. I have found that they are accomplices, but I am not at all sure that there was any degree of mens rea behind their actions. I find that they are accomplices because in the interests of the first accused and in the light of what they did I feel bound to do so. On the other hand I am abundantly satisfied that their evidence was truthful and therefore conclusive against the first accused."

The appeal to this Court is, first, on the ground that the learned Judge's decision constituted an error in law inasmuch as he reversed a finding of fact of the trial Magistrate and thereby avoided consideration of the grounds of the first appeal to which we have referred, and secondly on the ground that it was not open in law to the learned Magistrate to convict upon the uncorroborated testimony of persons who were found by him to be and who should accordingly be regarded by both appellate Courts as accomplices.

As to the first of those grounds, in our view the learned Judge was justified in the conclusion to which he came, expressed by him as follows:-

"The Magistrate's reasoning is a little difficult to follow, but as I understand his findings as to the pointsmen he regarded them merely as helping the thief in fact but without guilty knowledge. From such a finding I see no reason to dissent. On it the Magistrate should not have held them to be accomplices.'

We read the learned Magistrate's judgment in the same way as did the learned Judge. In effect the former, after finding the facts from which the finding of accomplice or no accomplice must be inferred, drew an inference which was excluded by the very facts which he had found. The learned Judge was not reversing a finding of the trial Court which depended to any extent on credibility, but was merely substituting the right inference for the wrong one which the Magistrate had drawn. He was correcting an error which it was open to him to correct. On second appeal to this Court the appellant is entitled as a matter of law to argue that the first appellate Court had no sufficient legal reasons to reverse an inference drawn from the facts as found at the trial. See Fazleabbas Sulemanji v. Reg., 22 E. A. C. A. 395. But, as we have said, we agree with the learned Judge's view. Accordingly the first ground of appeal to this Court fails.

It follows that the second ground of appeal to this Court does not arise, since the learned Magistrate's acceptance of the pointsmen's evidence as the truth must now be regarded as supplying ample corroboration of the accomplices Gulamhussein and Patel. Nevertheless, we think it right to record our view as to this second ground also, since it was submitted by counsel on both sides that it would be of much assistance if we were to put the matter beyond doubt.

In our opinion the position is as follows: $\rightarrow$

As has already been said by this Court in relation to appeals from other territories, the rules as to accomplices and corroboration laid down in *Davies* $v$ . Director of Public Prosecutions, (1954) A. C. 378 H. L. are binding on us. Every territory within our jurisdiction is in the same position in this respect. The fact that a Magistrate or Judge is sitting alone or with assessors neither excludes nor modifies the application of those rules. Provided that it is clear that the trial Court has expressly called to mind the necessary warning as to the danger of convicting on the uncorroborated testimony of an accomplice or of accomplices and has approached its decision on that footing, it is lawful to convict on such testimony if it is accepted as the truth beyond any reasonable doubt. We need hardly add that the danger in question is real and that the need fully and properly to appreciate that danger is no mere legal fiction or formality but a necessary practical safeguard which merits the most careful application in every case.

Mr. Dodd for the appellant submitted that in the East African Territories when a Judge or Magistrate sits alone or with assessors the Court should never convict in the circumstances under consideration, citing as his principal authority R. v. Haji Mohamed Saleh Mohamed, (1933) 15 L. R. Kenya 109, where this Court said in its judgment: —

"While we do not subscribe to the dictum that the corroboration must be overwhelming to sustain a conviction, our opinion is that in the absence of special or exceptional circumstances a conviction before a Judge or a Magistrate sitting alone, resting on uncorroborated accomplice evidence, is so dangerous that it should not be held."

In the same judgment the Court referred to its decision in Kichingeri v. R., (1908) 3 E. A. L. R. 1 where it upheld a conviction on uncorroborated accomplice evidence, the decision having been expressly arrived at "entirely on account of the exceptional circumstances". In our view any such reference by this Court to "special" or "exceptional" circumstances must be treated only as a rule of prudence, or as another mode of expressing the accepted principle as to the need to appreciate the danger of convicting in the type of case in question. As such, it is of course unexceptionable. But we cannot to-day accept Mr. Dodd's submission that it amounts to a rule of law peculiar to those territories from which appeals lie to this Court. The rules are, in our opinion, those enunciated in *Davies v. D. P. P.* and those rules are exhaustive.

It is, moreover, to be noted that any rule of law requiring "special" or "exceptional" circumstances to be proved in order to justify a conviction on uncorroborated accomplice evidence would, as seems sometimes to be overlooked, be inconsistent with section 133 of the Indian Evidence Act, 1872, which provides without exception or limitation that "a conviction is not illegal merely because it proceeds upon the uncorrborated testimony of an accomplice".

Mr. Dodd also cited R. v. Manilal Purohit, (1942) 9 E. A. C. A. 58. In that case this Court merely referred to the danger of convicting without corroboration and reaffirmed the principles as to what constitutes corroborative evidence. Nowhere in the judgment was it suggested that there must as a matter of law be "special" or "exceptional" circumstances to justify a conviction on accomplice evidence alone. What the Court did say was that corroboration of accomplice evidence "should be looked for", a proposition falling far short of Mr. Dodd's contention as to the rule of law, and one from which nobody would dissent.

There is, however, another decision of this Court, not cited on the hearing of this appeal, to which we should now refer, namely R. v. Thakar Singh $s/o$ Kahir Singh, (1934) 1 E. A. C. A. 110. The following passage, upon which the relevant part of the headnote is founded, appears in the judgment (at p. 112):-

"As regards corroboration a number of cases have been cited, but it seems necessary to refer only to the decision of the Court of Appeal for Eastern Africa in the case of Haji Mohammed Saleh (Cr. App. 108/1933) which settles the law applicable to this country to the effect that, save in exceptional cases, a conviction cannot be based upon the uncorroborated evidence of an accomplice. It is clear from that case and from the case of Kichingeri and others (3 E. A. L. R. 1) that the exception arises and the evidence of an accomplice without corroboration is sufficient, not in cases which are usually difficult of proof by other means, nor upon the trial of certain particular offences, but only when the circumstances of the case and the relation of the accomplice to the offence committed appear to justify the Court in acting upon his evidence."

$\overline{1}$

If this Court there intended to read R. v. Haji Mohamed Saleh's case (supra) as laying down a rule of law prohibiting a conviction on accomplice evidence alone "save in exceptional cases", the decision must be regarded as<br>not only irreconcilable with section 133 of the Indian Evidence Act but also as now overruled by Davies v. D. P. P. in which the opinion of their Lordships contained no suggestion of any such legal requirement. While adopting the latter part of the passage which we have cited from the judgment in $R$ . $v$ . Thakar Singh as a general indication as to when "the exception arises" in practice. we cannot accept the earlier part of that passage as to-day containing the true rule of law applicable to any of the territories with which this Court is concerned. R. v. Haji Mohamed Saleh must, as we have said, be treated not as having "settled the law" but as having expressed a rule of practical prudence to the effect that, save in exceptional cases, a conviction should not—not "cannot" -be based upon the uncorroborated evidence of an accomplice.

Next we should refer to R. v. Wamjerwa, (1944) 11 E. A. C. A. 93, cited by Mr. Summerfield for the Crown-respondent. This Court there held that, since the witness concerned was "only an accomplice in a very secondary sense" and was actuated by "relatively innocent motives", corroboration of her evidence was, though desirable, not essential, there being "as in *Kichingeri's case* exceptional circumstances which would justify a departure from the general rule". The appeal in that case was, however, dismissed not only upon that ground but also because it appeared to the Court that there was in fact some corroboration (upon which the trial Court had not relied), namely certain items of circumstantial evidence which the appellate Court described as sufficient "to furnish some slight corroboration in a case where the degree of criminal complicity to be attributed to the alleged accomplice is very slight indeed".

We agree with counsel that that case was one in which this Court was prepared to uphold conviction on the uncorroborated testimony of an accomplice, although the decision did not eventually rest on that view alone. The expression "exceptional circumstances", which once again appears in the judgment, should again be treated as indicative of no more than the rule of prudence to which we have referred.

It remains to be observed that this is not the first time that this Court has expressed the view which we take of the practical desirability of, as opposed to the legal necessity for, evidence corroborating that of an accomplice. In R. v. Ndaria and others, (1945) 12 E. A. C. A. 84 at p. 86, the following passage appears in the judgment—given, be it noted, at a later date than that of any of the other judgments of this Court to which we have referred:—

"If at the trial A gives evidence, such evidence is on the same footing as that of any other witness and may be taken into consideration against B, irrespective of whether A implicates himself or not. B, of course, can crossexamine and the question of accomplice evidence again has to be borne in mind in deciding how far it is safe to act upon this evidence as against B.

In cases in which the confession or evidence of A can be taken into consideration against B, it will, generally speaking, be considered unsafe to convict $B$ on such accomplice evidence, unless there is some independent corroboration tending to implicate B. On this the leading case is, of course, Rex v. Baskerville, (1916), 2 K. B. 658, which should always be referred to in case of difficulty."

The words "it will, generally speaking, be considered unsafe to convict" of course constitute the essence of the matter for present purposes, bearing in mind that the case was one in which the trial was before a Judge sitting with assessors and not with a jury. It is true that in the passage cited this Court was considering the case of persons (A and B) jointly charged, with the result that the hypothetical accomplice A would not be a Crown witness but one called in his own defence. Nevertheless it appears from the words "such evidence is on the same footing as that of any other witness" that the subsequent observation as to it being "generally speaking considered unsafe to convict" is to be taken as applicable to a case where the accomplice testifies for the Crown.

Our conclusion on this whole question may be summarized as follows: The rules of law as to what persons are to be treated as accomplices and as to corroboration of their evidence are those laid down in Davies v. D. P. P. Those rules apply equally to a trial with or without a jury; where there is no jury the Court must warn itself, and should also warn the assessors, if any, of the danger of convicting on accomplice evidence alone. The danger is real and the judge or judges of fact must be fully conscious of its reality. Generally speaking it is the practice, founded upon prudence when applying the rule as to the onus of proof, not to convict without any evidence corroborating that of accomplices. But there are exceptional cases in which a departure from that general practice is justified. The criterion as to whether such an exceptional case has arisen is the credibility of the accomplice or accomplices combined with the weight to be attributed to the facts to which they testify. The principal factors to be considered when assessing their credibility are not only their demeanour and quality as witnesses but also their relation to the offence charged and the parts which they played in connexion therewith, that is to say, the degree of their criminal complicity in law and in fact. A departure from the general rule of practice is only justifiable where, on applying that criterion in that manner, it clearly appears that the accomplice evidence is so exceptionally cogent as to satisfy the Court beyond reasonable doubt, and where accordingly the judge or judges of fact, while fully conscious of the general inherent danger of any such departure, is or are convinced that in the particular instance concerned the danger has disappeared.

The appeal is dismissed.