In the 1954--1960 period, Kilauea Volcano, Hawaii, erupted four times: summit eruptions occurred in 1954 (Halemaumau) and 1959 (Kilauea Iki) and flank eruptions occurred in the lower East rift zone in 1955 and 1960. Our analysis demonstrates that the chemical compositions of lavas from each eruption are related solely through fractionation and sorting of the observed phenocryst assemblages. We have used Pearce element ratio diagrams to make the following observations and inferences: 1) The conserved element ratio P/K for each eruption has less variance than is expected from analytical uncertainty. Thus, we cannot reject the hypothesis: Each set of lavas represents a comagmatic suite of compositions. 2) Where each of the four groups of lava analyses is plotted on a Pearce element ratio diagram with axes of Si/K and <0.5(Mg+Fe)+1.5Ca+2.75Na+0.25Al>/K, the data fit a slope of 1.0. Therefore, we cannot reject the hypothesis: The chemical variation in each data set is consistent with sorting of OL¿PL¿AU. 3) Each of the 4 groups of lavas has the same mean P/K ratio. As a result, we cannot reject the hypothesis: The 1954--1960 lavas originated from a single magma batch or several magmas with identical P/K values. Finally, using the most magnesian glass analysis in the Kilauea data suite (1959: SG--05), we have modelled the differentiation path with thermodynamic-based calculations. These calculatins predict a liquid line of descent that is consistent with the path inferred from rock and glass analyses. Therefore, the simulations provide an additional test of our hypotheses and corroborate the postulated differentiation model. Additionally, these calculations suggest that differentiation took place at less than 0.3 Gpa and extended over a temperature range of between 100¿C and 150¿C. ¿ American Geophysical Union 1990