On the Universal Splitting Property

We prove that if an incomplete computably enumerable set has the the universal splitting property then it is low2. This solves a question from Ambos-Spies and Fejer [1] and Downey and Stob [7]. Some technical improvements are discussed.