The scintillation counting behavior of a group of fifty-five pure crystalline organic compounds has been extensively studied. The data obtained have been analyzed with the goal of developing a better understanding of the scintillation process and of scintillator behavior. The materials were carefully purified, and massive crystals were grown. Relative scintillation average-pulse-height efficiencies at 30°C and —70°C for cobalt-60 gamma-ray excitation, gamma-ray excited scintillation decay times, and 2537 A ultraviolet-excited reflection and transmission photofluorescence spectra have been determined. A few solutions were also studied for comparison purposes. The purification and properties of the different materials are discussed in detail. The experimental data have been analyzed on the basis of Birks' photon cascade theory of the scintillation process. The ratio of the scintillation efficiency to the integrated photofluorescence intensity is shown to be a measure of Birks' primary photon production efficiency. The primary photon efficiencies vary over a relatively narrow range, so that there is a rough general correlation between the scintillation efficiencies and the integrated photofluorescence intensities. The best phosphors are those with rigid molecular and crystal bonding so that both the primary photon production and photofluorescence efficiencies are reasonably high. There is a tendency for the best scintillators to have the simplest and most sharply defined photofluorescence spectra. The scintillation efficiencies have been correlated with molecular structures and the mobility of the π electrons (or with the resonance interactions) within the molecules. In the simpler cases, consideration of singly charged quinoid structures permits the calculation of parameters which can be directly correlated with the scintillation efficiencies. In more complex cases, steric hindrance, hyperconjugation, ``unshared'' and ``nonbonding'' electron pairs, five-membered aromatic rings, heavy atom (triplet-state) effects, and quinoid (quenching) ground states must be considered. In all cases, the effects of thermal (vibrational) perturbations are present. ``Bond density'' and ``bound valence'' values have proved to be useful in correlating scintillation phenomena and molecular structure effects. Strongly colored compounds are usually poor scintillators. Earlier predictions of high scintillation efficiency for quinquephenyl and sexiphenyl are reaffirmed, and coronene, benzo[ghi]-perylene, and 1,2,5,6-dibenzanthracene are also predicted to be significantly better than the presently known organic scintillators. The further study of compounds containing five-membered aromatic rings appears promising. A generally good correlation has been found between the scintillation efficiencies and molecular diamagnetic anisotropy values. Some relationships between the scintillation decay times and the molecular structures can also be seen. The best crystalline organic scintillators are colorless substances of high melting point possessing molecules of simple structure and low atomic number in which there is extensive resonance conjugation of rings, ethylenic double bonds, and other groups to give extended, rigidly interlocked systems.