This paper presents the first results from a survey of the integrated spectra of 90 nearby galaxies. Intermediate- and low-resolution spectrophotometry over the 3650-7000 angstrom range has been used to compile a spectral atlas of galaxies (published in a companion paper) and to investigate the systematic behavior of the emission-line spectra in normal and peculiar galaxies. The data are especially useful as a comparison sample for spectroscopic surveys of more distant galaxies. The integrated absorption- and emission-line spectra show a smooth progression with Hubble type. Most spiral and irregular galaxies exhibit detectable emission lines of H-alpha, [O II], [N II], [S II], and sometimes H-beta and [O III]. The H-alpha line is the strongest emission line and is most reliable quantitative tracer of the massive star formation rate, but the [O II] line provides a good substitute in high-redshift galaxies, where H-alpha is inaccessible. The H-beta and [O III] lines are unsatisfactory as star formation tracers in all but the strongest emission-line galaxies, due to variations in stellar absorption and excitation, respectively. Methods for distinguishing between distant starburst galaxies and active nuclei are discussed. The new data are used to examine the statiscal properties of the [O II] emission in galaxies, and to calibrate a mean relation between [O II] luminosity and the total star formation rate. Star formation rates derived from the [O II] line are less accurate than those derived from H-alpha, and comparable in accuracy to rates derived from the continuum fluxes and colors. However, the [O II] data are very useful for comparing the star formation properties of large samples of galaxies. The distribution of [O II] equivalent widths in our sample is in good agreement with previous surveys of nearby galaxies and very different from that observed for blue-selected galaxies at intermediate redshift. The [O II] distribution for faint galaxies is very similar to that which is observed for nearby Markarian galaxies.