Some land-risk stories do not end with a dramatic finding of fraud. This one is important because the Court of Appeal treated suspicion, conflicting searches, and a pleaded prohibition as insufficient without the register evidence needed to carry a fraud case.
In Selina Mecca Wekesa v Kennedy Ellam & 4 others, the Court of Appeal at Kisumu considered Civil Appeal No. 46 of 2014, arising from Bungoma HCCC No. 54 of 2006. The dispute concerned L.R. No. Ndivisi/Muchi/1086, later subdivided into Ndivisi/Muchi/6454 to 6463. The judgment was delivered on 3 April 2020 in a combined file containing J. Mohammed JA's main judgment and Hannah Okwengu JA's concurrence.
What the appeal was about
The appellant said she was entitled to protection over the land and challenged the subdivision and transfers that followed. She relied in part on an asserted prohibition associated with Mombasa HCCC No. 457 of 2002 (O.S.) and on conflicting land-search evidence. For this note, that Mombasa matter is treated only as evidentiary history described in the appeal, not as a separately reviewed source.
At the High Court, the trial judge had rejected the case against the 2nd, 3rd and 4th respondents, treating them as innocent purchasers for value without notice. The High Court also made findings that placed fraud liability on the 1st and 5th respondents. That distinction matters. The appellate question was not whether the story sounded suspicious. It was whether fraud and the pleaded register interference had been strictly proved on the record.
The Court of Appeal answered no. J. Mohammed JA emphasized that fraud must be distinctly pleaded and strictly proved. The appellant had not produced the underlying register, nor adequate proof that the alleged prohibition had been registered and later removed fraudulently. Conflicting search certificates and uncertainty in registry evidence could not, by themselves, establish collusion or fraudulent removal of a prohibition.
Why the result matters for land-risk review
For buyers, lenders, advocates, and family-asset decision makers, the case is a useful warning against relying on a single reassuring or alarming document. A prohibition, a subdivision history, and a transfer chain all need record-level support. Where the register itself is missing from the proof, a court may be unwilling to infer fraud from inconsistency alone.
The appeal was dismissed. The Court of Appeal left the 2nd, 3rd and 4th respondents' innocent-purchaser treatment undisturbed on this record, but that should not be read as a universal shield for every downstream buyer. The point is narrower: the pleaded fraud was not proved against them. The 5th respondent's cross-appeal was allowed because there was no legal basis to sustain the fraud finding against that respondent. Costs of the appeal went to the 2nd, 3rd, 4th and 5th respondents, and costs of the cross-appeal went to the 5th respondent. Okwengu JA concurred, including on the failure to prove fraud and the final orders.
ArdhiScore takeaway
This case shows why land diligence has to ask two questions at once: what does the paper trail appear to say, and what proof would survive challenge if the paper trail is contested? A buyer or lender who only sees a title number may miss an old dispute. A claimant who only sees suspicious movement may still fail without the underlying register, proof of entries, and a clear evidentiary chain.
ArdhiScore is built for that gap. It does not replace legal advice or declare current ownership, possession, rectification, succession, or later parcel status. It helps surface public judgment signals, parcel-linked dispute history, and risk context so users can ask better questions before money, credit, or family expectations are committed.
Source: Court of Appeal decisions - Civil delivered on April 3, 2020, member locator CoA - Civil - April 3, 2020/(Combined) Kisumu Civil Appeal 46 of 2014 Final.pdf.
Learn how the workflow turns judgment evidence into practical checks: How ArdhiScore works.