"Ultrareletivistic nuclear collisions at the Relativistic Heavy-Ion Collider have produced a high temperature, high energy density medium consisting of a strongly interacting plasma of quarks and gluons. This extreme state of matter provides a testing ground for quantum chromodynamics. Previous studies of gold-gold collisions over a wide range of beam energies revealed many properties of the produced medium. However, these studies were restricted to relatively large colliding systems which resulted in large collision volumes; it is therefore important to investigate what role the size of the collision volume plays in the evolution of the source, particularly as the source volume becomes vanishingly small. This can be achieved with symmetric copper-copper collisions, which offer access to a range of system sizes from [approximately] 10 participating nucleons up through volumes comparable to those created in gold-gold collisions. Collective behaviors of the produced particles in heavy-ion collisions can provide useful probes into the state of the medium produced, including its degree of thermalization and its properties. The elliptic flow, an anisotropy in the azimuthal distribution of the produced particles that is strongly correlated to the initial transverse geometry of the colliding nuclei, is one such collective motion that has proven to be a very useful observable for studying heavy-ion collisions. This is because it exhibits fairly large magnitudes in the systems being studied and is sensitive to the strength of the partonic interactions in-medium. The PHOBOS experiment, which can measure the positions of produced charged particles with high precision over nearly the full solid angle, is well-suited to study the elliptic flow and its evolution over an extended range along the beam direction. The elliptic flow from copper-copper collisions at center-of-mass energies of 22.4, 62.4, and 200GeV per nucleon pair are presented as a function of pseudorapidity and system size. The appearance of unexpected behaviors in the smaller system prompted a re-examination of the role of the collision geometry on the production of elliptic flow. Studies using Monte-Carlo Glauber simulations found that the fluctuating spatial configurations of the component nucleons in the colliding nuclei could result in significant variation of the shape of the nuclear overlap on an event-by-event basis, and that these fluctuations become important for small systems. The eccentricity, a quantity that characterizes the ellipticity of the nuclear overlap in the transverse plane, is redefined to account for these fluctuations as the participant eccentricity. It is found that the event-by-event fluctuations of the participant eccentricity are able to fully account for the observed elliptic flow in the smaller system. The participant eccentricity is used to normalize the measured elliptic flow across different colliding systems to a common initial geometry so that a direct comparison of the properties of the produced medium can be made. It is found that the produced medium evolves smoothly from systems of [approximately] 10 participant nucleons to systems involving more than 350 nucleons and for collision energies from 19.6 to 200GeV per nucleon pair. This smooth evolution of the elliptic flow is also observed as a function of pseudorapidity in all the systems studied. After accounting for the initial geometry, no indication of the identity of the original colliding system is observed"--Page vi-vii.