Eight characteristics of sediments are considered as possible means of recognizing source beds: 1, quantity of organic matter in the sediments; 2, reducing power, which is a measure of ability of the sediments to reduce chromic acid; 3, color of sediments; 4, volatility of sediments; 5, degree of volatility, which is a measure of the volatility with respect to the organic content; 6, ratio of carbon to nitrogen in the sediments; 7, oxidation factor, which is an index of the state of oxidation of the sediments, and is the ratio of the carbon content to' the reducing power; and, 8, the nitrogen-reduction ratio, which is the ratio of the nitrogen content to the reducing power. Several thousand determinations of these characteristics have been made and averaged from more than 800 lithologic units from many areas in the California, Rocky Mountain, Mid Continent, East Texas, and Gulf Coast oil regions. The thickness of these units ranges from 50 ft. to 500 ft. Each of them has been classified according to its probable richness in source material of petroleum. Nearness to producing zones has been the basis of classification, as it seems more reasonable to assume that oil in general accumulates near where it is generated than to assume that it ordinarily accumulates far from where it is made. Three classes were made: 1, lithologic units within 250 ft. stratigraphically above or below an oil zone and less than 2 miles from an oil field; 2, (a) units within 500 ft. above or below an oil zone and within 15 miles of an oil field (excluding, of course, those that belong to the first class) and (b) units within 250 ft. stratigraphically of a horizon and more than 15 miles distant from an oil field, provided the unit is located within a region in which the horizon is generally productive of oil or yields significant quantity of oily substances when extracted with ether; and, 3, all other beds, viz., those more than 500 ft. above or below an oil horizon or more than 15 miles distant from an oil field and beds at greater distance if they are within 250 ft. of horizons that generally contain oil in the area. These three classes are designated in this report as " productive," " questionably productive," and " barren." " Productive " units naturally may contain spine beds that are poor in source material, and " barren " units may contain beds that are rich in source material; but, on the whole, the " productive " units-because of their nearness to oil zones-are more likely' to be richer in mother substances of petroleum than are the " barren " units. Each of the eight characteristics was averaged for each of the three classes of productivity, for each of the five oil areas: California, Rocky Mountains, Mid Continent, East Texas, and Gulf Coast. The average organic content of the "productive" and " barren " units was found to be approximately the same in each region studied, from which it is inferred that the quantity of organic matter in a sediment probably is not a reliable guide of the ability of the sediments to generate oil. The average reducing power of the " productive " beds is slightly greater than that of the "barren"; but the difference is so slight that the reducing power, also, probably is not a satisfactory guide to source beds. The color of, the sediments, in general, becomes darker as the organic content of the sediments increases; and, as the organic content seems to be a poor index of source beds, color presumably is not a serviceable means of recognizing source beds. The volatility of the " productive " beds is greater than that of the " barren " units in each of the five regions, but it differs so much from region to region that it cannot be used as an index of source beds unless the general volatility of the sediments in the region under consideration is known. For example, the average volatility of the " barren " units in California and the Rocky Mountain regions is greater than that of the " productive " units in the Mid Continent, East Texas, and Gulf Coast areas. The degree of volatility differs less from region to region than does the volatility, and in general exhibits a relationship to productivity equally as good as the volatility. The average ratio of carbon to nitrogen in each of the three regions from which data are available-Rocky Mountains, Mid Continent, and East Texas-is higher for " productive " beds than for " barren " units, but the ratio differs somewhat from one region to another. The oxidation factor of the " productive " units is approximately the same as that of the " barren " units in the Rocky Mountain region and in the East Texas region, but not in the Mid Continent region-where the oxidation factor of the " productive " units is definitely lower than that of the " barren " units. The average nitrogen-reduction ratio of the " productive " units is decidedly greater than that of the " barren " units in each of the five regions studied; and it, therefore, is the most reliable of the guides studied. Relatively few " productive " units have nitrogen-reduction ratios greater than 7.0, and relatively few " barren " units have ratios of less than 4.5. A considerable number of both " productive " and " barren " units have ratios ranging between 4.5 and 7.0, but even between these limits the productive units tend to be more commonly associated with low ratios than are the " barren " units. Four of these eight characteristics-volatility, degree of volatility, carbon-nitrogen ratio, and nitrogen-reduction ratio-therefore, give promise of being useful means of recognizing source beds; and one of them, the nitrogen-reduction ratio, is particularly encouraging. Additional studies of these four characteristics, and especially of the nitrogen-reduction ratio, are highly desirable.