According to these reports on more than two dozen tests, in-situ combustion, when property engineered, can recover 50 percent of oil in place. However, lest you be fired up, in the heat of statistics, to ignite your reservoir, bear in mind that man by nature is disinclined to expose his failures. Introduction In-situ combustion may be said to date back to 1888 when Mendeleev suggested the in-situ conversion of coal into combustible gases. In 1934, Sheinman and Dubrovai, on the basis of earlier laboratory studies, proposed the process of oil recovery by means of a proposed the process of oil recovery by means of a moving underground firefront. Following this, a number of field tests were carried out in Azerbaidjan and other regions from 1934 to 1937. An analysis of the tests showed that the heat losses were so large that injected hot gases reached the formation face with zero thermal energy. Nevertheless. the tests opened the way to more field tests, laboratory research, and theoretical studies, as a result of which in-situ combustion has come to be recognized as a promising method of recovering oil. At the present time there are nearly 300 papers, and close to a dozen books (in Russian) dealing partly or wholly with in-situ combustion. At least nine large-scale field in-situ combustion projects are currently in operation. Listings of field projects are currently in operation. Listings of field tests can be found in a number of publications (for example, Refs. 5, 6, 51, 54, 61, 64, 65). The main purpose of this paper is to present, compare, and contrast the performance of two dozen forward combustion tests that have been reported in detail out of 100 or so that have been cited in the literature. Reverse combustion and various modifications of the basic process will not be considered. Only brief mention will be made of wet combustion tests. Also, the paper is not intended to give criteria for designing a combustion project, nor does it purport to discuss equipment and operational aspects of firefloods, except for the special operational problems encountered in fireflooding. No reference will be made to a great many experimental and mathematical studies conducted principally in the U. S., U.S.S.R., and Holland; however, a few instances will be cited where the experimental and predicted results were compared with the test results. Principle Statistics of Selected Firefloods Principle Statistics of Selected Firefloods Tables 1A and 1B summarize the principal statistics of 24 forward combustion tests. These, and the two cited below. seem to be the only ones for which information is available on oil recovery or air-oil ratio (AOR), or both. In a few cases, certain performance data were estimated from the reported performance data were estimated from the reported data. Such values are indicated by a question mark. It is recognized that the highly complex in-situ combustion process, compounded by reservoir heterogeneities, cannot be expressed in terms of over-all parameters. However, in spite of its limitations the table should be of value in balancing out the most favorable conditions for a fireflood. Notice that the table does not give values of the producing gas-oil ratio (GOR) water-oil ratio (WOR), producing gas-oil ratio (GOR) water-oil ratio (WOR), ignition information, gas recovery, produced gas composition, sweep data, etc. These items are covered in separate sections. The test numbers used in the text refer to the corresponding tests in the tables. To complete the table, two tests should be added. The available information on them is rather meager, but the key data are available. JPT P. 477