Springs constitute an important source of water in east Tennessee, and many individual springs are capable of supplying the large quantities needed for municipal and industrial supplies. Most of the springs in east Tennessee issue from solution openings and fractured and faulted zones in limestone and dolomite of the Knox Group, Chickamauga Limestone, and Conasauga Group. The ability of these rocks to yield a sustained flow of water to springs is dependent on a system of interconnected openings through which water can infiltrate from the land surface and move to points of natural discharge. Ninety springs were selected for detailed study, and 84 of these are analyzed in terms of magnitude and variability of discharge. Of the 84 springs analyzed, 4 flow at an average rate of 10 to 100 cfs (cubic feet per second), 62 at an average rate of 1 to 10 cfs, and 18 at an average rate of 1 cfs or less. Of the 90 springs, 75 are variable in their discharge; that is, the ratio of their fluctuations to their average discharges exceeds 100 percent. Mathematical analysis of the flow recession curve of Mill Spring near Jefferson City shows that the hydrologic system contributing to the flow of the spring has an effective capacity of about 70 million cubic feet of water. The rate of depletion of this volume of water, in the absence of significant precipitation, averages 0.0056 cfs per day between the time when the hydrologic system is full and the time when the spring ceases to flow. From such a curve it is possible to determine at any time the residual volume of water remaining in the system and the expected rate of decrease in discharge from that time to cessation of flow. Correlation of discharge measurements of 22 springs with those of Mill Spring shows that rough approximations of discharge can be projected for springs for which few measurements are available. Seventeen of the springs analyzed in this manner show good correlation with Mill Spring: that is, their coefficients of correlation were 0.70 or better as compared with a perfect correlation factor of 1.00.