We use spin-density-functional theory to study the spacing between conductance peaks and the ground-state spin of 2D model quantum dots with up to 200 electrons. Distributions for different ranges of electron number are obtained in both symmetric and asymmetric potentials. The even/odd effect is pronounced for small symmetric dots but vanishes for large asymmetric ones, suggesting substantially stronger interaction effects than expected. The fraction of high-spin ground states is remarkably large. PACS numbers: 73.23.Hk, 73.40.Gk, 73.63.Kv The interplay of quantum mechanical interference and electron-electron interactions is a current theme in many areas of solid-state physics: the 2D metal-insulator transition, interaction corrections in mesoscopic systems, and efforts toward solid-state quantum computing, for instance. A semiconductor quantum dot (QD) [1, 2] -a nano-device in which electron motion is quantized in all three dimensions -is a particularly simple system in which to study this interplay. In Coulomb blockade experiments in the electron tunneling regime, the conductance through the dot varies strongly as a function of gate voltage, forming a series of sharp peaks. For closed dots at low temperature, both the positions and heights of the peaks encode information about the dot's ground state. In particular, the spacing between adjacent conductance peaks is proportional to the second difference of the ground state energy with respect to electron number N , ∆ 2 E(N ) ≡ E gs (N + 1)+ E gs (N − 1) − 2E gs (N ), which is often called the addition energy. Furthermore, the ground state spin of the QD can be inferred from the shift in position of the conductance peaks upon applying a magnetic field.The addition energy varies because of changing interference conditions either as N changes or from dot to dot, leading to a conductance-peak-spacing distribution. Previous theoretical work addressing this distribution can be divided into roughly two types: First, computational approaches addressed small dots with randomly disordered potentials -both exact diagonalization [3,4,5] and self-consistent field methods 7,8,9,10] or density functional theory [11,12]). Second, a semi-analytic treatment of large dots was developed based on general statistical assumptions [2,13,14,15,16,17]: An important contribution to the variation comes from the single-particle energy; to treat this for irregular quantum dots, one assumes that the single-particle dynamics is classically chaotic, and so the single-particle quantum properties can be described by random matrix theory (RMT) [2,18]. A random-phase approximation (RPA) treatment of the screened electron-electron interaction was then combined with such an RMT description of single-particle states to describe dots with large N .The results of these two approaches are quite different.First, for the zero temperature peak-spacing, the small dot calculations yield Gaussian-like distributions while the large N results are non-Gaussian. Second, spin degeneracy causes a significant "eve...