Eight Argonne Premium Coal samples and two other coal samples were used to observe the effects of combustion and leaching on 30 elements. The results were used to infer the modes of occurrence of these elements. Instrumental neutron activation analysis indicates that the effects of combustion and leaching on many elements varied markedly among the samples. As much as 90% of the selenium and bromine is volatilized from the bituminous coal samples, but substantially less is volatilized from the low-rank coals. We interpret the combustion and leaching behavior of these elements to indicate that they are associated with the organic fraction. Sodium, although nonvolatile, is ion-exchangeable in most samples, particularly in the low-rank coal samples where it is likely to be associated with the organic constituents. Potassium is primarily in an ion-exchangeable form in the Wypdak coal but is in HF-soluble phases (probably silicates) in most other samples. Cesium is in an unidentified HNO3-soluble phase in most samples. Virtually all the strontium and barium in the low-rank coal samples is removed by NH4OAc followed by HCl, indicating that these elements probably occur in both organic and inorganic phases. Most tungsten and tantalum are in insoluble phases, perhaps as oxides or in organic association. Hafnium is generally insoluble, but as much as 65% is HF soluble, perhaps due to the presence of very fine grained or metamict zircon. We interpret the leaching behavior of uranium to indicate its occurrence in chelates and its association with silicates and with zircon. Most of the rare-earth elements (REE) and thorium appear to be associated with phosphates. Differences in textural relationships may account for some of the differences in leaching behavior of the REE among samples. Zinc occurs predominantly in sphalerite. Either the remaining elements occur in several different modes of occurrence (scandium, iron), or the leaching data are equivocal (arsenic, antimony, chromium, cobalt, and nickel). The results of these combustion and leaching experiments indicate that some previously held assumptions concerning modes of occurrence of elements in coal should be reconsidered.