ar X iv : h ep - p h / 01 06 07 5 v 1 7 J un 2 00 1 MRST GLOBAL FIT UPDATE

We discuss the impact of the most recent data on the MRST global analysis-in particular the new high-E T jet data and their implications for the gluon and the new small x structure function data. In the light of these new data we also consider the uncertainty in predictions for physical quantities depending on parton distributions, concentrating on the W cross-section at hadron colliders. There has recently been a great deal of updated data which help to determine parton distributions, particularly on small x structure functions from HERA 1,2 and on inclusive jets from the Tevatron 3,4. In both cases these new data are both more precise and extend the kinematic range, and thus determine the parton distributions more accurately than ever before. The MRST global fit is performed by minimizing the χ 2 for all data adding statistical and systematic errors in quadrature (this procedure will be improved before producing finalized distributions), except for the jet data where the dominant correlated errors render a proper treatment absolutely essential. The evolution begins at Q 2 0 = 1GeV 2 , and data are cut for Q 2 < 2GeV 2 and W 2 < 10GeV 2 in order to exclude regions where higher twist and/or higher orders in α S are expected to play an important role. The main effect of the new data is to constrain the gluon distribution, and hence, via evolution the small x quark distributions. The jet data are important in determining the high x gluon. Previously 5,6 MRST used prompt photon data to perform this role, but it has become clear that there are both large theoretical uncertainties and conflicts between data sets, and we now drop this approach. The best fit to jet data produces a high x gluon in between our previous g and g ↑ distributions, i.e. xg(x) ∼ (1 − x) 5.5. If the gluon becomes smaller at high x the fit to the jets becomes unacceptable. If it becomes larger, then although the jet data are initially supportive, the fit to