Petrography of the Lower Ordovician Ellenburger Group, both in deeply-buried subsurface cores and in outcrops which have never been deeply buried, documents live generations of dolomite, three generations of microquartz chert, and one generation of megaquartz. Regional periods of karstification serve to subdivide the dolomite into "early-stage", which predates pre-Middle Ordovician karstification, and "late-stage", which postdates pre-Middle Ordovician karstification and predates pre-Permian karstification. Approximately 10% of the dolomite in the Ellenburger Group is "late-stage". The earliest generation of late-stage dolomite, Dolomite-L1, is interpreted as a precursor to regional Dolomite-L2. L1 has been replaced by L2 and has similar trace element, O, C, and Sr isotopic signatures, and similar cathodoluminescence and backscattered electron images. It is possible to differentiate L1 from L2 only where cross-cutting relationships with chert are observed. Replacement Dolomite-L2 is associated with the grainstone, subarkose, and mixed carbonate-siliciclastic facies, and with karst breccias. The distribution of L2 is related to porosity and permeability which focused the flow of reactive fluids within the Ellenburger. Fluid inclusion data from megaquartz, interpreted to be cogenetic with Dolomite-L2, yield a mean temperature of homogenization of 85 + or - 6 degrees C. On the basis of temperature/delta 18 O water plots, temperatures of dolomitization ranged from approximately 60 to 110 degrees C. Given estimates of maximum burial of the Ellenburger Group, these temperatures cannot be due to burial alone and are interpreted to be the result of migration of hot fluids into the area. A contour map of delta 18 O from replacement Dolomite-L2 suggests a regional trend consistent with derivation of fluids from the Ouachita Orogenic Belt. The timing and direction of fluid migration associated with the Ouachita Orogeny are consistent with the timing and distribution of late-stage dolomite. Post-dating Dolomite-L2 are two generations of dolomite cement (C1 and C2) that are most abundant in karst breccias and are also associated with fractures, subarkoses and grainstones. 87 Sr/ 86 Sr data from L2, C1, and C2 suggest rock-buffering relative to Sr within Dolomite-L2 (and a retention of a Lower Ordovician seawater signature), while cements C1 and C2 became increasingly radiogenic. It is hypothesized that reactive fluids were Pennsylvanian pore fluids derived from basinal siliciclastics. The precipitating fluid evolved relative to 87 Sr/ 86 Sr from an initial Pennsylvanian seawater signature to radiogenic values; this evolution is due to increasing temperature and a concomitant evolution in pore-water geochemistry in the dominantly siliciclastic Pennsylvanian section. A possible source of Mg for late-stage dolomite is interpreted to be from the dissolution of early-stage dolomite by reactive basinal fluids.