20100509 TZCA Mwanza

1 IN THE COURT OF APPEAL OF TANZANIA

ATMWANZA (CORAM: MSOFFE, J.,A, MBAROUK, J.,A And BWANA, J.A) CRIMINAL APPEAL NO. 220 OF 2005 G ER_AID LUCAS ......................................... ..................... APPELLANT

VERSUS THE REPUBLIC ............................................................ RESPONDENT (Appeal from the Judgment of the High Court of Tanzania at Mwanza) (Rweyemamu, J.) dated the 6th day of June, 2005 in Criminal ADDeaI No. 1169 of 2000

JUDGEMENT OF THE COURT 4th & 111h May, 2010 BWANA,J.A.: The Appellant, Gerald Lucas, was one of four persons charged with two counts of Armed Robbery contrary to sections 285 and 286 of the Penal Code. The other three were acquitted by the trial court (District Court of Mwanza at Mwanza). The appellant was convicted of the two counts and sentenced to serve two concurrent terms of thirty (30) years of imprisonment. Aggrieved by that decision he appealed before the High Court, an appeal which was unsuccessful. He has now preferred this second appeal. 1 Before the trial court, the prosecution alleged that on 20th October, 2000 at 10.30pm, armed bandits broke into the shop of one Samwel Simon (PWI) and stole an assortment of articles valued at shs. 108,4501=. Some minutes later, the same bandits did break into a room occupied by Damas Paul (PW2) and stole one sewing machine valued at shs. 70,0001=. In both incidents, the bandits did use actual violence to the two people mentioned above by hitting them with an iron bar. PW1 was hit on the right hand in order to obtain the stolen property while PW2 was hit on his head for the same purpose. Six prosecution witnesses testified for the prosecution while four testified for the defence. Both Pwl and PW2 claimed to have identified the appellant as being one of the bandits who broke into the house at that night hour. According to PW1, it is

the appellant who held him on the floor and hit him with the iron bar. It was PW1's further evidence that he knew the appellant because the latter lived along the same street before moving away. On his part, PW2 testified that he identified the appellant as one of the four bandits who broke into his room at that late hour of the night. Both PW1 and PW2 were subsequently taken to hospital by patrol police who arrived at the scene of the robbery shortly, responding to an alarm raised. The police did not, however, manage to arrest the bandits but the name of the appellant was mentioned by PW1, as being one of the bandits who had robbed him. He was arrested some days later and charged accordingly, together with - three other suspects. - - Although totally denying responsibility to the two counts, the appellant did admit that he once lived in the vicinity of where pwl and PW2 live - at Mahina Ihogo area within the city of Mwanza. He had moved away to Nyakato before the robbery incident. Both the trial and the first appellate courts found the appellant guilty of the offences on the basis of identification both visUal and voice. The courts "a quo" arrived at that conclusion after evaluating the evidence on identification. According to the judgment of the High Court, it did not base its decision on the cautioned statement which is 3 why the other co-accused persons implicated under the

said statement, were acquitted. The issue of adequate identification forms the basis of this appeal. The appellant insists that he was not adequately identified at the scene of crime. However, Mr. Victor Kahangwa, learned Senior State Attorney, representing the respondent Republic, was of the view that the identification of the appellant was watertight. We are, therefore, of the considered view that this appeal can be disposed of by considering the issue of identification that is, whether or not the appellant was adequately identified at the scene of crime. Both the trial court and the first appellate court did analyze the factual issues involved and came to the settled conclusions that led to the conviction of the appellant. This being a second appellate court, we should not-as per long settled principle - willingly interfere with those factual findings unless (but not restricted to) there are glaring errors on the face of the record, errors in calculations, or misdirections or non directions on 11 the evidence (see DPP vs Jaffari Mfaume Kawawa (1981) TLR 149; Issa Mgara @ Shuka vs Republic - Criminal Appeal No. 37 of 2005 - unreported; Paschal Christopher vs DPP - Criminal Appeal No. 106 of 2006 (unreported). In Edwin Mhondo vs Republic (1993) TLR 170 at 174, this Court held:- "On a second appeal to this Court, we are only supposed to deal with questions of law. But this

approach rests on the premise that the findings of fact are based on a correct appreciation of the evidence. If, as in this case, both courts completely misapprehended the substance, nature and quality of the evidence, resulting in an unfair conviction, this court must, in the interest of justice, intervene." We now turn to the issue of identification - both visual and voice-of the appellant at the scene of crime. The crime with which the appellant was convicted of took place at around 10:30pm. The prosecution case, as stated earlier, relies on the evidence of PW1 and PW2 for identifying the appellant. We firstly consider visual 5 0 identification. Were the conditions adequate and correct for identification? According to the evidence of PW1 and PW2, both were asleep in their respective rooms. Then bandits broke in and ordered to be given money. Pwl testified that he identified the appellant because there was light. He did not specify what kind of light it was. However, according to PW2, he could identify the appellant by the help of "the light of the beam of the torch" from the other bandits. If this evidence is to be relied upon, then serious doubts do arise. This is so because, as this Court stated in James Chilonji vs Republic (Criminal Appeal No. 101 of 2003 - unreported); if PW2 daim was that it was the flashlight which enabled her to identify the appellant, that could not

be because a flash li'ht (torch) enables the person using it to see the object or person fit on and not the other way round' We subscribe to the foregoing view. PW2 could not identify the appellant using flashlight as that would have been directed towards him and not the other way round. There is no dearth of authorities governing visual identification. The then Court of Appeal of Eastern Africa did lay down a cardinal principle on the subject in the case of Abdullah bin Wendo and Another vs Rex (1953) 20 EACA 116. The views expressed in that case were adopted by this Court in the celebrated case of Waziri Amani vs Republic (1980) TLR 250. The principle (aid down in these cases is that in a case involving evidence of visual identification, no court should act on such evidence unless all possibilities of mistaken identity are eliminated and that the court is satisfied that the evidence before it is absolutely watertight A decision to be made in respect of visual identification should be in compliance with what was stated in Raymond Francis vs Republic (1994) TLR 100, 103 namely:- ". It is elementary that in a crIminal case where determination depends essentially on identification, evidence on conditions favoring a 7 correct identification is of the utmost importance (See also Musa Abdullah vs Republic - Criminal Appeal No. 36 of 2005 - unreported; Rizal Rajabu vs Republic -

Criminal Appeal No. 110 of 2006, unreported; Paschal Christopher vs DPP - supra). All these authorities lead to the irresistible conclusion that identification must be watertight. In our respected view such watertight evidence is said to exist when it leads to the exclusion of all possibilities of mistaken identity. The court should therefore consider the following before arriving at its decision:- First - how long did the witness had the accused person under his/her observation. Second - What was the estimated distance between the two people. Third - If it were at night (as in the instant case) which kind of light did exist. Fourth - had the witness seen the accused person before the day and time of crime. If so when and how often. Fifth - the whole evidence before the court considered, are there material impediments or discrepancies affecting the correct identification of the accused by the witness. Sixth - in the course of the observation of the accused by the witness, was there any obstruction experienced by the witness, obstruction which may have interrupted the latter's concentration. The evidence on recOrd in this case does not meet the six guidelines stated above. Therefore we are hesitant to hold that the appellant was properly identified by either PW1 or PW2. That is not all. There is PW2's evidence that he could as well identify the appellant by his voice. Voice

identification is - as was stated by this Court in Stuart Erasto Yakobo vs Republic (Criminal Appeal No. 202 of 2004 - unreported): "... the issue is whether voice identification is reliable in law. In our considered opinion voice identification is one of the weakest kind of evidence and great care and caution must be 66 taken before acting on it ... there is a/ways a possibility that a person may imitate another person's voice. For voice identification to be relied upon it must be established that the witness is very familiar with the voice in question as being the same voice of a person at the scene of crime (Emphasis added). (See also Badwin Komba @ Ballo vs Republic- Criminal Appeal No. 56 of 2003 - unreported; Kanganja Ally and Juma Ally vs Republic - [1980] TLR 270). Familiarity of an accused's voice has many dimensions. The accused may be related to the witness as a blood relative; may have lived or stayed together for quite some time, and the like. The list of such possibilities is never restricted. In the instant case, however, we do not see anything to the effect that PW2 had previous long acquaintances with the appellant so as to enable the former recognize the latter's voice unmistakably and even to the exclusion of imitation. The above considered, we are hesitant to uphold the conviction based on both visual and voice identification of the appellant. The 10 said

identification appears to be problematic and does not withstand the tests enumerated above. As a consequence of that, we think it is unsafe to uphold the conviction of the appellant. We accordingly allow the appeal, quash the conviction and set aside the sentence of thirty (30) years imposed on the appellant by the trial court and upheld by the first appellate court (the High Court). We do order that unless the appellant is otherwise lawfully held, he should be set free forthwith. DATED at MWANZA, this gth day of May, 2010. .J. H. M SOFFE

JUSTICE OF APPEAL M. S. MBAROUK

JUSTICE OF APPEAL S.

3. BWANA

JUSTICE OF APPEAL I certify isisa true copy of the original. IN ' 0 1\ '(J1I \ ikG IA) 11 \