ZIMBABWE: PARLIAMENTARY ELECTIONS - 2005 16 trends in the various provinces and constituencies". The discrepancies do not prove that the final figures are wrong, but the differences are so high that they call for an explanation. Such an explanation is still pending. The official results are closer to the turnout of 2002, than to the results as per 2 p. m., and it is possible that the discrepancy is due to incomplete reporting and not engineering of results. A clarification of this can only be reached by full access to the results tabulation. In most countries, polling stations results are made public by the elections authorities, in order that observers and the general public may confirm the accuracy of the tabulation. In Zimbabwe making detailed results public would be the only way of removing all the suspicions expressed after the elections. The parties and the observers, in particular a professional organisation such as ZESN, should have been able to publicise parallel counts based upon their own polling station data immediately after the count. Only on 22 April did ZESN finalise their final analysis. The analysis is based upon observer reports from a selection of polling stations in 115 out of the 120 constituencies. It adds up the results as reported by the observers for each constituency from those polling stations where they were present. This selection of polling station results is added up to constituency level and the percentage of support for each candidate is compared with the official results as made public by the ZEC. ZESN concludes that the ZEC results in favour of ZANU PF ‘vary with the ZESN pattern in four constituencies, Chipinge South, Buhera South, Makoni East and Gwanda’. These are constituencies where ZESN feels they have sufficient data to draw such a conclusion. However, ZESN’s sample even in these constituencies is around 25% of the votes, except for in Gwanda where it is 50% (of the ZEC figures). It is therefore difficult to draw any final conclusions. ZESN’s selection of data need not be representative for the constituency. There will often be large variations in results within a constituency and it is not a given that the polling stations where ZESN observers were permitted to observe the count are representative. On the other hand, for the four constituencies, the data do form a solid basis for raising questions. This makes the ZEC’s hesitation to publish the polling station tabulation counterproductive, and it underpins the need for full transparency in the calculation of results. The Effects of the Delimitation of Constituencies A full assessment regarding whether undue political considerations were used would only be possible to perform with detailed local knowledge. The following is based upon overall figures only. The Harare South constituency was a new one. This was the only one won by ZANU PF in Harare. The constituency did not follow natural boundaries, and seemed to be constructed with a political purpose in mind. Manyame was a new constituency in Mashonaland West. The town Norton was transferred into Manyame from Mhondoro. Mhondoro changed from MDC to a big ZANU PF majority, and Manyame was won by ZANU PF. Mhondoro had after the