The Taconic rock sequence extends from near Sudbury, Vermont, to near Poughkeepsie, New York, a length of about 150 miles; and from just west of the Green Mountain Range and Berkshire Highlands to the valleys of the Hudson River and Lake Champlain, a width of about 20 miles. The Taconic rocks are now in the axial region of the Middlebury synclinorium and its southward extension. The Taconic sequence consists of about 2000 feet of slate, with subsidiary graywacke, quartzite, and limestone; many of the rock units are turbidites. Fossils from the rocks include forms of Early, Middle, and Late Cambrian and Early and Middle Ordovician age. Evidence of stratigraphic tops derived from the fossils and from primary sedimentary features agrees with structural data and demonstrates that the Taconic rocks are geometrically the highest strata within the Middlebury synclinorium. The underlying rocks of the synclinorium (the synclinorium sequence) are right side up; they also range in age from Early Cambrian to Middle Ordovician but belong to a different sedimentary lithofacies (dominantly carbonate and orthoquartzite). The relation between the Taconic sequence and the synclinorium sequence, therefore, is a baffling problem. Lithostratigraphically, the Taconic sequence falls into three groups: (1) the pre- Normanskill "low Taconic" sequence, occurring in the area between the main Taconic Range and the Hudson River, as far south as Rhinebeck, New York; (2) the Normanskill Shale in the same area, as well as in the area south of Rhinebeck at least as far as Poughkeepsie, New York, and also west of the Hudson River; and (3) the "high Taconic" sequence, occupying the main Taconic Range from Dorset Mountain, Vermont, south to Indian Mountain in Sharon, Connecticut, as well as Mount Greylock in Massachusetts. Rocks of (1) and (2) are fossiliferous, but to date no fossil has been found in (3). Rocks of (1) and that part of (2) areally coextensive with (1) thus are of known age but uncertain three-dimensional geometric configuration, whereas rocks of (3) are of known configuration (in the centers of open synclinoria) but unknown age. Only that part of (2) beyond the areal confines of (1) is both of known age and known configuration; these rocks are in sedimentary contact above the older rocks of the synclinorium sequence and are autochthonous. At the north end of the Taconic sequence in western Vermont, rocks of group (1) are beyond reasonable doubt allochthonous. Because of the southward geometric continuity of the structural elements, all the Taconic rocks of group (1), and that part of group (2) areally coextensive with it, are interpreted as allochthonous. The structure of group (3), the high Taconic sequence, is inferred by topography and by detailed lithostratigraphic matching with rocks of the east Vermont sequence; on this basis, as well as on the basis of the broad lithic similarity with rocks of group (1), rocks of group (3) are concluded to be also allochthonous. A discontinuous polymict conglomerate underlies and surrounds the allochthon on all sides and is interpreted here as a record that dates the imminent arrival of the allochthon at each locality. The conglomerate contains unsorted blocks of rocks of both the Taconic sequence and the synclinorium sequence; the matrix is the autochthonous upper Normanskill Shale or its equivalent. Fossils from the matrix shale date the event as Trenton, probably Sherman Fall in age. The geologic history of the area is reconstructed as follows: The pre-Normanskill Taconic rocks were deposited in the area of the present Precambrian massifs of the Green Mountains-Berkshire Highlands belt between the clastic, eugeosynclinal east Vermont sequence to the east and the miogeosynclinal synclinorium sequence to the west; they constitute the transitional facies between these two belts. Conditions were relatively stable until early Middle Ordovician time, when the Green Mountain- Berkshire Highlands area began to rise and the area of the present Middlebury synclinorium began to subside. Subsidence took place largely by a series of high-angle longitudinal faults that, as a whole, step down to the west. Argillaceous sediments (the Normanskill Shale) began to inundate the former miogeosynclinal area; because the conditions of sedimentation had become similar, the sediments resembled, in facies, the synchronous Taconic rocks that were being deposited to the east. Continued rise of the Green Mountains-Berkshire Highlands area led in middle Trenton time to the decollement of the Cambrian and Ordovician sediments into the area of the present Middlebury synclinorium in a series of giant submarine slides. Sedimentation continued at the receiving site throughout the event; sedimentation may also have persisted on the moving slides. The record is found today in the turbidite-laden shale and graywacke in the upper part of the Normanskill Shale of both the allochthon and the autochthon. Restoration of the allochthonous rocks to the original site of deposition leads to correlations between rocks of the Taconic sequence and of the largely autochthonous east Vermont sequence. The lithic correlation can be carried to the level of individual formations and is confirmed by a few known ages in the east Vermont sequence. Several lines of reasoning lead to a plausible correlation of part of the Cavendish Formation of southeastern Vermont with the oldest part of the Taconic sequence. This correlation leads further to the conclusion that in this area the contacts between the Green Mountain massif and the Cavendish Formation and between the Cavendish and the overlying east Vermont sequence must both be thrust faults of large displacements. This conclusion is in fact inevitable because one of the Taconic thrust slices that extends without interruption between the latitudes corresponding to the gap in the Precambrian massifs has been shown by local structural evidence to be allochthonous; an outside original depositional site must be found for it. The present Taconic allochthon is coextensive with an area of marked negative Bouguer gravity anomaly; the Green Mountains-Berkshire massifs constitute a belt of positive anomaly. It is here proposed that these anomalies resulted from a deepseated transfer of material; subcrustal addition of material caused the rise of the Green Mountains-Berkshire Highlands area, and the concurrent subtraction of material caused subsidence in the Middlebury synclinorium area through a series of faults which were the near-surface expression of an episode of crustal collapse. If this interpretation is correct, then the regional gravity anomaly represents an uncompensated feature that has persisted since Middle Ordovician time.