Analysis of deep inelastic scattering with z-dependent scale

Evolution of the parton densities at NLO in αS using W̃ 2 = Q2(1 − z)/z instead of the usual Q2 for the scale of the running coupling αS is investigated. While this renormalisation scale change was originally proposed as the relevant one for x → 1, we explore the consequences for all x with this choice. While it leads to no improvement to the description of DIS data, the nature of the gluon at low x, low Q2 is different, avoiding the need for a ‘valence-like’ gluon. 1. Motivation − Non-singlet evolution In deep inelastic scattering (DIS) the Q dependence of flavour non-singlet quantities is quite straightforward. Taking a non-singlet structure function, e.g. F NS = F ep 2 − F en 2 or F NS = F ν̄p 2 − F νp 2 then at next-to-leading order (NLO) we have F NS(x, t) = q̃ ± NS(x, t) + ( αS 4π ) Bq(z)⊗ q̃ NS(x/z, t) (1) where t = lnQ , q̃(x, t) = xq(x, t), the relevant combination of quark and antiquark densities (weighted by the appropriate charges squared) is denoted by qNS and Bq(z) is the quark coefficient function given, for example, in the MS scheme by