In Volume 1 of this report, results of Canada-Ontario phosphorus removal treatability studies were summarized and regression equations for predicting chemical requirements for phosphorus removal based on influent phosphorus levels were developed. The second phase of the study, as reported herein, was carried out to determine whether improved prediction equations could be derived by considering several wastewater characteristics, in addition to influent phosphorus. To this end, a multi-parameter jar testing program was carried out on 20 different raw municipal wastewaters. A wastewater strength index (WSI) was developed which was used to rank and classify wastewaters by simultaneously considering six common parameters (hardness, total phosphorus, suspended solids, total alkalinity, conductivity and total organic carbon). Multiple regression relationships, expressing alum, iron salt and lime dosages as a function of raw wastewater characteristics were derived for three wastewater strength (weak, medium, strong) categories. The equations obtained for lime were inconsistent and had little predictive value. However, a good linear regression relationship between alkalinity and lime required to reach a given pH was developed. Jar tests, in conjunction with such a relationship, remain the best way to estimate lime requirement for a given wastewater. The predictive value of the Al and Fe equations was somewhat better than that of the simpler equations developed in the first phase of the study, based on the influent phosphorus concentration only (i.e.,± 20% compared to = ± 30%). However, this degree of improvement would not warrant the extra time and expense normally required for sample collection and analysis for several wastewater characteristics. Unless the additional information required for these equations is readily available, the simpler equations developed in Volume 1 should be used to assist in the selection of the chemical and dosage for full scale application.