africa.lawBeta
SearchAsk AICollectionsJudgesCompareMemo
africa.law

Free access to African legal information. Legislation, case law, and regulatory documents from across the continent.

Resources

  • Legislation
  • Gazettes
  • Jurisdictions

Developers

  • API Documentation
  • Bulk Downloads
  • Data Sources
  • GitHub

Company

  • About
  • Contact
  • Terms of Use
  • Privacy Policy

Jurisdictions

  • All jurisdictions →

© 2026 africa.law by Bhala. Open legal information for Africa.

Aggregating legal information from official government publications and public legal databases across the continent.

Back to search
Case Law[2021] TZCA 470Tanzania

Lucas s/o Evance @ Bwandu & Another vs Republic (Criminal Appeal 392 of 2018) [2021] TZCA 470 (16 September 2021)

Court of Appeal of Tanzania

Judgment

IN THE COURT OF APPEAL OF TANZANIA AT MBEYA (CORAM: NDIKA. J.A.. SEHEL, J.A., And KENTE. J.A.^ CRIMINAL APPEAL NO. 392 OF 2018

  1. LUCAS S/O VENANCE @ BWANDU]
  2. GODFREY S/O BARNABA J7 ...... VERSUS T -- APPELLANTS THE REPUBLIC RESPONDENT (Appeal from the Judgment of the High Court of Tanzania at Sumbawanga) 13th & 16th September, 2021 NDIKA, J.A.: On 5th November, 2018, the High Court of Tanzania sitting at Sumbawanga (Mambi, J.) convicted Lucas s/o Venance @ Bwandu and Godfrey s/o Barnaba, the first and second appellants respectively, of murder and sentenced them to death. By this appeal, the appellants now challenge their convictions. At the trial, it was common ground that Marko s/o Peter ("the deceased") met a violent death on 4th April, 2014 at Kakese area within Mpanda District and Region of Katavi. What was hotly contested was whether the appellants were the perpetrators of the crime. (Mambi, J.) dated the 5th day of November, 2018 in Criminal Sessions Case No. 23 of 2016 JUDGM ENT OF THE COURT

To establish its case against the appellants, the prosecution relied upon the testimonies of five witnesses. The evidence tended to show that three youths aged between 13 and 15 years, namely, Paulo Peter (PW1), Shabani Mikaeli (PW2) and Edward Charles (PW3) along with the deceased and one other person, went fishing in River Mpanda on 4th April, 2014. While on the land along the edge of the river around 19:00 hours, two young men, each holding a lighted torch raided them. It was claimed that with the aid of the light from the assailants' torches as well as moonlight, they saw the assailants and heard their voices. They recognized them as the appellants with whom they were familiar as they came from the same locality. Soon thereafter, the raiders started beating them. The witnesses escaped along with the other person, leaving behind two containers filled with their day's catch as well as the deceased who was overwhelmed by the raiders. As they ran away, they heard the deceased scream in anguish, "Mamaa, Mamaa . " Having escaped unscathed, the witnesses went home and slept. On the following morning, they confronted the appellants whom they found selling fish but they refused to give them back. There and then, they reported the matter to PW4 Afla Mwanandenje, a local leader. PW4 confronted the appellants who then claimed to have spotted the deceased

hobbling around earlier that day. Later that day, the deceased's body was recovered from the river. PW5 Emmanuel Akisa, a Clinician at Mpanda Health Centre, conducted an autopsy on the deceased's body. He issued a post-mortem examination report (Exhibit P.l), declaring that the deceased, a child of the tender age of 13 years, died owing "to head injury due to assault by a blunt o b je ct " The appellants interposed the defences of a lib i and general denial. Apart from stating that they were not at the scene of the crime on the fateful night, they recounted on the manner of their arrest and denied generally to have been familiar with the prosecution witnesses. Following the learned trial Judge's summing up at the conclusion of the cases for the prosecution and defence, two of the assessors who sat with him returned guilty verdicts against the appellants while the other assessor opined that the offence was unproven. Siding with the first two assessors, the learned trial Judge went ahead and convicted the appellants of murder and condemned them to death as hinted earlier. Briefly, in his judgment, the learned trial Judge, having reviewed the evidence on record in the light of the relevant case law on visual identification, held that, based on the eyewitness evidence of PW1, PW2 3

and PW3, the appellants were positively recognised at the scene as the assailants who attacked the deceased. While acknowledging that torchlight might have not aided identification, he based his finding on recognition by sight and voice upon the evidence that the scene was brightly moonlit and that the witnesses heard the appellants following an exchange of utterances between them. He went on to hold that the appellants must have been the perpetrators of the murder on the ground that they were the persons last seen with the deceased alive in view of their failure to give any exculpatory explanation. Finally, the learned Judge weighed the appellants' a lib is and general denials but rejected them. For the appellants, Mr. Daniel Lawrence Muya, learned counsel, lodged a memorandum of appeal in terms of Rule 73 (2) of the Tanzania Court of Appeal Rules, 2009 in substitution of the earlier memoranda of appeal lodged separately by the appellants on 4th January, 2019. Three grounds of complaint have thus been cited:

  1. That the tria i Judge erred in law in holding that the appellants were properly identified a t the scene o f the crime.

  2. That the learned tria l Judge erred in law to refer to extraneous evidence which was com pletely opposed to the testim onies o f PW1, PW2 and PW3.

  3. That the learned tria l Judge erred in law by holding that circum stantial evidence adduced by PW1, PW2 and PW3 irresistibly points to the appellants' g u ilt as they were the last persons to be seen with the deceased. In his oral argument in support of the appeal, Mr. Muya appraised the evidence on record in support of the prosecution case and urged us to find that the conditions at the scene were not conducive for a positive identification of the assailants on the following reasons: first, that if the identification of the assailants was aided by moonlight and torchlight, it was not stated if the torchlight illuminated the assailants' faces. Secondly, if the assailants used torchlight at the scene, it was most probably that the alleged moonlight at the scene was too weak. Thirdly, that the witnesses did not see the appellants assault and kill the deceased as they fled the scene shortly after the raid. Fourthly, that none of the witnesses identified the assailants' attire. Fifthly, that the witnesses' unusual delay in reporting the incident to the village authorities until the next day raised doubts over their credibility. Finally, that the utterances allegedly made by the appellants that led to their recognition by voice were not disclosed by the witnesses.

Citing W aziri Am an v. Republic [1980] TLR 250 as well as the unreported decisions of the Court in Said Chaly Scania v. Republic, Criminal Appeal No. 69 of 2005; Jum apili M syete v. Republic, Criminal Appeal No. 110 of 2014; and Yusuph Sayi & 2 O thers v. Republic, Criminal Appeal No. 589 of 2017, Mr. Muya submitted that, the judicial approach in identification evidence cases cautions against reliance upon such evidence without eliminating all possibilities of mistaken identification. In the instant case, he submitted, the evidence was too weak for a positive identification. Coming to the complaint that the learned trial Judge relied upon extraneous matters in his reasoning, Mr. Muya referred us to the questioned judgment, at page 91 of the record, where the learned Judge indicated that it was in the evidence that the three prosecution witnesses had torches at the scene which were lighted towards the other side. Relying upon Shija s/o Sosom a v. D.P.P, Criminal Appeal No. 327 of 2017; Monde Chibunde @ Ndishi v. Republic, Criminal Appeal No. 328 of 2017; and Richard O tieno @ Gullo v. Republic, Criminal Appeal No. 367 of 2018 (all unreported) where the Court censured trial courts for including in their judgments facts which are not reflected in the recorded evidence in the proceedings, the learned counsel contended that the 6

learned Judge's finding of guilt was overly swayed by such an extraneous matter. Rounding off with the grievance in the third ground of appeal, Mr. Muya submitted that in view of the doubtful circumstances in which the appellants were allegedly recognized at the scene, it was not inferable that the appellants were the last persons to be seen with the deceased alive. To bolster his submission, he cited the case of Alkadi W illiam @ Supa v. Republic, Criminal Appeal No. 188 of 2005 where the Court followed the decisions of the Court of Appeal for Eastern Africa in Kasaja s /o Tibagina v. R. (1952) 19 EACA 268 and R. v. Siprian s/o Nshange (1947) 14 EACA 72 for the principle that circumstantial evidence must irresistibly lead to the conclusion of guilt as opposed to innocence. Replying for the respondent, Ms. Scholastica Ansgar Lugongo, learned Senior State Attorney, stoutly resisted the appeal. She submitted, based on the testimonies of the identifying witnesses, that the scene was well illuminated by torchlight and moonbeam for the witnesses to see and recognize the appellants with whom they were familiar. Citing PW2's evidence, she said that the assailants were also recognized by their voices. She added that the fact that the identifying witnesses reported the incident the following day to PW4 gave credence to their version of the events as

opposed to effacing their credibility. It was her contention that the delay might have been unusual but the three youthful witnesses could not have been expected to act more responsibly and conventionally. In response to the Court's probing, the learned Senior State Attorney conceded that the duration of the attack by the assailants was not stated and that the words allegedly exchanged at the scene were not stated by PW2. Regarding the contention in the second ground of appeal, Ms. Lugongo conceded unreservedly that the learned Judge considered extraneous matters in his judgment. However, she disagreed that the said extraneous matters clouded the learned Judge's reasoning and findings. On the final ground of appeal, the learned Senior State Attorney argued that it was in the evidence that after the identifying witnesses had fled the scene, the deceased remained in the hands of the appellants and that he was never seen alive since then. She added that the appellants told PW4 that they saw the deceased hobbling around the next day after the fateful incident but later that day his corpse was recovered from the river. On that basis, she claimed that the doctrine of last seen was rightly applied and urged us to dismiss the appeal. In a brief rejoinder, Mr. Muya reiterated his earlier submissions and prayed that the appeal be allowed.

We have examined the record of appeal and keenly considered the contending submissions of the learned counsel as well as the authorities cited. In our view, the sticking issues in the appeal are whether the appellants were positively identified at the scene as the assailants and if yes, whether the doctrine of last seen was justifiably invoked and applied against the appellants to found conviction. Ahead of dealing with the above issues, we find it convenient to deal with the grievance in the second ground of appeal that extraneous matters blighted so materially the learned trial Judge's reasoning and findings. Having reviewed the record of appeal, at page 91, we agree with the learned counsel that the learned trial Judge, indeed, slipped into palpable error by considering facts not borne out of the record. The impugned passage in the judgment reads thus: "The record shows that both sides (accused persons) and the three witnesses (PW1, PW2 and PW3) had torches and at one tim e each side lighted towards the other side . " The record clearly shows that the three identifying witnesses had no torches and that apart from moonshine their identification of the assailants was aided by light emitted from the assailants'torches. This anomaly apart, we noted two further perturbing instances in which extraneous matters

were considered in the judgment. While the identifying witnesses adduced that they fled the scene soon after the raid and heard the deceased scream in agony as they were running away, the learned trial Judge, at the same page 91, observed that: "These three witnesses managed to see the two accused very clearly attacking the deceased on that day where there was moonlight. "[Emphasis added] It is too plain for argument that the witnesses did not tell the trial court that they saw the appellants attacking the deceased. The learned trial Judge went on in his judgment, at pages 92 and 93 of the record, observing that: "... the fact that the accused persons came closer to PW1, PW2 and PW3 and exchanged some words in a bright light a ii these make this court believe that the identification was proper." [Emphasis added] The above observation is yet another aberration. None of the three eyewitnesses told the trial court about the proximity between them and the assailants nor did they say anything about the brightness of the illumination at the scene. Apart from PW1 stating, as shown at page 10, that soon after 10

the raid the attackers commanded them to run away, there was no evidence of any exchange of words between the two sides at the scene. As hinted earlier, Mr. Muya cited Shija s/o Sosom a {supra) and Monde Chibunde @ Ndishi {supra) and urged us to follow them. In those cases, we warned trial courts against including in their judgments facts which are not reflected in the recorded evidence in the proceedings. In Shija s/o Sosom a {supra), we followed our earlier decision in Athanas Ju lias v. Republic, Criminal Appeal No. 498 of 2015 (unreported) where we held the act of the trial resident magistrate to include in his judgment, facts which are not reflected in the record an incurable irregularity on the following reasoning: "The im plication here is that\ either, in his judgm ent ■ the tria l resident m agistrate did include extraneous m atters which did not com pletely feature in the evidence o f the witnesses who were called to testify, or, the tria l resident m agistrate did om it to record a num ber o f facts that were said by the witnesses in their testim onies. In either case, we are inclined to jo in hands with the contention o f the learned counsel for both sides that ■ the irregularity occasioned was fatal and did vitiate the entire proceedings o f the tria l court . " 11

In the instant case, we hesitate to travel the whole distance with Mr. Muya and hold the error at hand an incurable irregularity. We think the circumstances in the present case are somewhat different. The learned trial Judge's errors appear to be a result of misapprehension of the evidence on record but the integrity of the recorded evidence is beyond reproach. At any rate, however, we agree with Mr. Muya that the misapprehended facts clouded the learned trial Judge's reasoning and findings on the crucial issue whether the appellants were positively identified at the scene. Accordingly, we find merit in the second ground of appeal. We now deal with the issue whether the appellants were positively identified at the scene as the assailants. Since it is undisputed that the incident in the instant case occurred at night around 19:00 hours, the evidence on how the assailants were seen and recognized at the scene is so decisive. In its ground-breaking decision in W aziri Am an {supra), the Court cautioned, at pages 251 to 252, that the evidence of visual identification is of the weakest kind and most unreliable and that it should not be acted upon "unless all possibilities of mistaken identity are eliminated and the court is fully satisfied that the evidence before it is absolutely watertight." The Court stated further, at p. 252, that: 12

"Although no hard and fast rules can be laid down as to the manner a tria l Judge should determ ine questions o f disputed identity, it seems dear to us that he could not be said to have properly resolved the issue unless there is shown on the record a careful and considered analysis o f a ll the surrounding circum stances o f the crim e being tried. We would, for example, expect to find on record questions as the following posed and resolved by him: the time the witness had the accused under observation; the distance at which he observed him; the conditions in which such observation occurred, for instance, whether it was day or night-time, whether there was good or poor lighting at the scene; and further whether the witness knew or had seen the accused before or not These m atters are but a few o f the m atters to which the tria l Judge shouid direct his m ind before coming to any definite conclusion on the issue o f identity . " [Emphasis added] In Said Chaly Scania {supra), it was stressed that visual identification evidence must be clear and complete: "We think that where a witness is testifying about identifying another person in unfavourable circum stances, like during the night, he m ust give

dear evidence which leaves no doubt that the identification is correct and reliable. To do so, he w ill need to mention a ll the aids to unmistaken identification like proxim ity to the person being identified, the source o f lig h t and its intensity, the length o f time the person being identified was within view and also whether the person is fam iliar or a stranger." See also Raym ond Francis v. Republic [1994] TLR 100 for the principle that evidence on conditions favouring a correct identification is of utmost importance where identification is an issue; Ju m ap ili Msyete {supra) and Frank Maganga v. Republic, Criminal Appeal No. 93 of 2018 (unreported). In the instant case, the three identifying witnesses claimed that the appellants were familiar to them and that they saw and recognized them at the scene with the aid of moonlight and the light emitted from the assailants' torches. PW2 added another dimension to the evidence; he averred that he also recognized the appellants by their voices with which he was familiar. However, on the whole, we are of the view that the misgivings expressed by Mr. Muya over the cogency of the identification evidence are unassailable. For a start, we agree that, if moonlight sufficiently illuminated 14

the scene, it would have negated the use of torchlight. It is inferable from the averred use of torchlight that either moonlight was too weak to aid proper identification or there was none of it. It should be recalled that none of the witnesses addressed the trial court on the intensity of the moonlight. It is also likely that light emitted from the assailants' torches directly to the identifying witnesses could not have effectively aided them to see and recognize the assailants. As we held in Moham ed M usero v. Republic [1993] TLR 290, beams of torches from thieves tend to dazzle and impair the vision of the witnesses into whose eyes they are directed. Moreover, there was no detail as to how long the witnesses observed the assailants before they fled the scene. Our impression from the evidence is that the witnesses fled the scene frantically shortly after the raid, implying that they had little or no time to observe their attackers. Turning to the claim that PW2 recognised the appellants by their voices, we would, at first, recall our view in Nuhu Selem ani v. Republic [1984] TLR 93 at page 94, that "it is notorious that voice identification by its e lf is not very r e lia b le In the case at hand, we have no doubt that PW2's claim was equally of little probative value, if any. Mr. Muya is correct that PW2 did not state the utterances made by the appellants by which he recognized their voices. In the frantic situation at the scene, it was unlikely 15

that there was any meaningful exchange of words between the two sides for a positive identification to have been made based on voice. We find it apt to recall our observation in Moham ed Musero {supra) at page 293 when we confronted an akin situation: "With regard to the voice this was also m ost unreliable in the circum stances o f this case. There was not much exchange o f words in this confused atmosphere, only one word 'tu lia' seem s to have been uttered and possibly another two lete pesa' when the bandits were demanding money. This to us appears insufficient to enable the witnesses to make a dear identification based on uwce. "[Emphasis added] We also endorse Mr. Muya's submission that the credibility of the identifying witnesses was questionable. It was strange that they went home and slept as if nothing unusual had befallen them. It was also outlandish that the first thing they did on the following day was confronting the appellants demanding the return of the fish allegedly seized from them the previous night. Only after the appellants had allegedly rebuffed their demand that they reported the alleged raid to PW4. In M arw a W angiti & Another v. R epublic [2002] TLR 39 at page 43, the Court observed that: "The a b ility o f a witness to name a suspect a t the earliest opportunity is an all-im portant assurance o f 16

his reliability, in the same way as unexplained delay or com plete failure to do so should put a prudent Court to inquiry." So far as it relates to visual identification, the Court held in Jaribu Abdalla v. R epublic [2003] TLR 271 at page 273 that: "In m atters o f identification, it is not enough m erely to look a t facts favouring accurate identification, equally im portant is the credibility o f the witness. The a b ility o f the witness to name the offender at the earliest possible moment is a reassuring ; though not a decisive factor." Ms. Lugongo sought to explain the delay by contending that the three witnesses were too young to determine the proper course of conduct. With respect, we disagree. In our considered view, at their age, between 13 and 15 years, they were mature enough and that they should have reported the matter that fateful night to the relevant authorities. They must have appreciated that they had left the deceased alone in a perilous situation and that he needed help. Their odd inaction renders their version fanciful and doubtful. That said, we hold that the appellants were not positively recognised at the scene. In the premises, we find merit in the first ground of appeal. 17

Given our foregoing finding that the appellants were not recognized at the scene, the doctrine of last seen was inapplicable to their case. The second issue is, accordingly, resolved in the negative. We thus find merit in the third ground of appeal. In sum, we hold that the offence of murder against the appellants was not proven beyond reasonable doubt against the appellants. On that basis, we allow the appeal and proceed to quash the convictions and set aside the sentence of death imposed on the appellants. Consequently, we order that the appellants, Lucas s/o Venance @ Bwandu and Godfrey s/o Barnaba, be released from prison unless they are detained there for other lawful cause. DATED at M BEYA this 15th day of September, 2021 G. A. M. NDIKA JUSTICE OF APPEAL B. M. A. SEHEL JUSTICE OF APPEAL P. M. KENTE JUSTICE OF APPEAL This Judgment delivered this 16th day of September, 2021 in the presence of the Appellants in person and represented by Mr. Daniel Lawrence Muya, learned counsel and Ms. Irene Mwabeza, learned State Attorney for the Respondent/Rejmklic, is hereby certifie s a true copy of the original.

Discussion