"The purpose of this volume is to gather the latest experiment results from the H1, ZEUS and HERMES collaborations and to capture new trends in HERA phenomenology. The presentations are by experts for experts, but are suitable for a mixed readership of both theoreticians and experimentalists. H1 members also cover ZEUS results and vice versa. This is the place where discrepancies between experimental data and theoretical predictions are pointed out and ventilated and where projects to be launched in the future are identified."--BOOK JACKET.

The purpose of this volume is to gather the latest experimental results from the H1, ZEUS and HERMES collaborations and to capture new trends in HERA phenomenology. The presentations are by experts for experts, but are suitable for a mixed readership of both theoreticians and experimentalists. H1 members also cover ZEUS results and vice versa. This is the place where discrepancies between experimental data and theoretical predictions are pointed out and ventilated and where projects to be launched in the future are identified.

One of the main approaches toward understanding the hadronic structure is the introduction of transverse momentum dependent parton distribution functions (TMDs). These non-perturbative objects are accessible through deeply inelastic processes such as semi-inclusive deep inelastic scattering (SIDIS), Drell-Yan and e+e- annihilation. The current thesis explores the hadronic structure within the TMD approach under three independent chapters. As the first study in the present thesis the complete cross-section for the production of unpolarized hadrons in e-p semi-inclusive deep-inelastic scattering is computed up to power-suppressed O(1=Q2) terms, utilizing an approximation which systematically assumes that ̄qgq-terms are much smaller than ̄qq-correlators. All leading- and subleading-twist structure functions are calculated in terms of six TMDs and two fragmentation functions for which phenomenological extractions are available. The results are compared to available experimental data from COMPASS, HERMES and Jeerson Lab. p Drell-Yan process is the subject of the second study in this thesis. All leading-twist structure functions of -induced Drell-Yan with polarized protons are computed at the scale of COMPASS experiment. The non-perturbative inputs for the TMDs are taken from available phenomenological extractions plus two well-established models: The light-front constituent quark model and the spectator model. Evolution eects are estimated using a Gaussian ansatz with energy-dependent Gaussian widths. The results are compared to COMPASS data when available leading to important conclusions about pion and proton TMDs. Predictions are made for future experiments. The third study in this thesis aims at establishing a generalized covariant parton model that allows us to obtain all possible TMDs in a systematic way by describing the quark correlator in terms of two covariant distribution functions for on-shell partons. The proposed formulation of the model successfully reproduces all results of the original version of the model for the leading-twist time-reversal-even (T-even) TMDs and enables us to evaluate all T-even twist-3 TMDs for the first time. The unpolarized and helicity parton distribution functions are used as inputs and other TMDs are calculated then compared to available TMD parametrizations. The results show several model relations between the TMDs some of which are expected from QCD and some verified in other quark models as well.