Scientific core drilling in the Pleistocene Valles caldera complex (encompassing the Valles (1.13 Ma) and coaxial Toledo (1.50 Ma) calderas) of north central New Mexico has provided new insight into the origins of sandstones, breccias, and pyroclastic surge deposits interbedded with the thick intracaldera ignimbrite sequence. These rocks were previously interpreted from geothermal drill cuttings as dominantly fluvial in origin. As such, representing significant erosional intervals, they formed much of the basis for subdividing the intracaldera ignimbrite sequence (up to 2000 m in apparent thickness where drilled) into four major units: the lower tuffs (>1.50 Ma); the Otowi (1.50 Ma) and Tshirege (1.13 Ma) members of the Bandelier Tuff; and a new unit, the upper tuffs, believed to be post-Bandelier in age (<1.13 Ma). All but the upper tuffs correspond to mapped outflow facies ignimbrite sheets. However, Continental Scientific Drilling Program core holes VC-2A (completed in 1986) and VC-2B (completed in 1988), in the Sulphur Springs area of the Valles caldera, have provided continuous cores, revealing for the first time that some intracaldera rocks previously though to be exclusively clastic actually have multiple origins. Some of these rocks are probably pyroclastic surge deposits; others could be lithic-rich breccias of various origins incorporated nearly instantaneously in ignimbrites during ash flow eruption and concomitant caldera collapse. These new findings demonstrate the value of continuous core for subsurface characterization and correlation of complex intracaldera lithologies; they also necessitate revision of Nielson and Hulen's (1984) cuttings-based intracaldera stratigraphic framework. For example, the hematitic S2 ''sandstone'' was initially interpreted as marking an erosional interval between the Tshirege Member of the Bandelier Tuff and the overlying, petrographically similar upper tuffs. Core from VC-2A and VC-2B, however, shows that the S2 cuttings could also represent disaggregated.

Permian red bed-rich, lithic lag breccias or caldera collapse mesobreccias. If this is the case, then most or all of the upper tuffs are actually uppermost Tshirege Member ignimbrites. In similar fashion and upon review of previously applied correlation criteria the ''lower tuffs'' of the western Valles caldera complex could represent both genuine pre-Bandelier ignimbrites and those of the lowermost Otowi Member. The core, however, shows that in the Sulphur Spring subsurface the lower tuffs are separated from overlying ignimbrite sheets by prominent erosional and eruptive breaks; they appear to be slightly more mafic than the overlying tuffs and host distinctive pumice lapilli. At this site the lower tuffs almost certainly predate the Otowi Member and are probably correlative with the outflow facies San DiegoCanyon ignimbrites (1.78 Ma). Cores from VC-2A and VC-2B support earlier interpretation of the S3 ''sandstone'' as a major marker horizon separating the intracaldera Otowi and Tshirege members of the Bandelier Tuff but clearly shows that this important unit is not, as previously thought, entirely a simple intracaldera epiclastic apron. In VC-2A the S3 has the superficial appearance of a sandstone but contains abundant blocky shards as well as a accretionary and armored lapilli; it is also soft-sediment-deformed and invades overlying nonwelded tuff as small clastic dikes. We believe that here the S3 was emplaced by a wet pyroclastic surge. In nearby corehole VC-2B the S3 consists of a basal, massive, sediment-gravity-flow(?) sandstone overlain by sandstone and dacite breccias with accretionary and armored lapilli-bearing tuffaceous matrices. These deposits are probably caldera collapse mesobreccias that were formed simultaneously with early Tshirege Member ash flow eruptions through or into a Toledo caldera lake. ¿ American Geophysical Union 1991