We investigate how probe density influences hybridization for unlabeled target oligonucleotides that contain mismatched sequences or targets that access different binding locations on the immobilized probe. We find strong probe density effects influencing not only the efficiency of hybridization but also the kinetics of capture. Probe surfaces are used repeatedly, and the potentially large contributions of sample-to-sample variations in surface heterogeneity and nonspecific adsorption are addressed. Results of kinetic, equilibrium, and temperature-dependent studies, obtained using in-situ surface plasmon resonance (SPR) spectroscopy, show that hybridization for surface immobilized DNA is quite different from the well-studied solution-phase reaction. Surface hybridization depends strongly on the target sequence and probe density. Much of the data can be explained by the presence of steric crowding at high probe density; however, the behavior of mismatched sequences cannot be understood using standard models of hybridization even at the lowest density studied. In addition to unusual capture kinetics observed for the mismatched targets, we find that the binding isotherms can be fit only if a heterogeneous model is used. For mismatched targets, the Sips model adequately describes probe-target binding isotherms; for perfectly matched targets, the Langmuir model can be used.