environmental conditions. There are three and probably four zones of calcium sulphate deposits (gypsum and anhydrite), averaging roughly 50 feet each, separated by varying amounts of brick-red argillaceous shale and fossiliferous limestone with subordinate thin dolomitic sandy shales, oölites, and calcareous algal bands. In amount, the gypsum and anhydrite may make up some 10 per cent of the total estimated thickness of 1100 feet, with red shale 50 per cent and calcareous beds 40 per cent. Estimates of the thicknesses of this series, however, must be considered as but very rough approximations, owing to the poorness and isolated nature of the exposures, and to the presence of numerous fault slips of which mention has been made. The sequence of faunal zones presented here is necessarily dependent on the writer's interpretation of the structure, and is not advanced as an established fact. Better and less broken sections may be found elsewhere in the province that will necessitate a considerable revision of the following succession. The sequence as given below is radically different from that presented by Hartt (1867) and in Dawson's "Acadian Geology." Upper Windsor or Martinia zone. Characterized by

Martinia opertacosta n. sp.

Subzone D. Characterized by Caninia dawsoni (Lambe) 460'± Subzone C. Characterized by Dibunophyllum lambei n.

sp.

Lower Windsor or Composita zone.
Composita dawsoni (Hall & Clarke).

270'

Characterized by

265'

Subzone B. Characterized by Diodoceras avonense (Dawson)

Subzone A. Basal limestone and Sanguinolites phase.. 95'

The Windsor beds disconformably overlie the Cheverie formation. The basal member is a fine calcareous quartzite whose upper surface is locally marked by interior molds of the valves of a Sanguinolites. Commonly both valves are attached and flatly opened. A thin conglomeratic development with limestone pebbles, probably of an intraformational character, follows the basal arenaceous bed, and is succeeded by platy brecciated limestone which carries calcite vugs and occasional nodules of pyrolusite and manganite of secondary origin. This manganiferous mineralization is of such widespread occurrence at this horizon that it is probable that the manganese content was originally disseminated in the limestone at the time

of deposition. The subzone terminates with a thick band of anhydrite and gypsum.

Subzone B is characterized by the presence of twolimestones that are extremely fossiliferous, and as they outcrop in the vicinity of the town of Windsor, they afford a rich collecting ground. The lower or Maxner limestone is about 80 feet in thickness, the upper or Miller limestone about 35 feet, the two being separated by a second band of gypsum. The fauna is common to the two limestones, but the association is somewhat different. The Maxner limestone is characterized particularly by the abundance of Diodoceras aronense in its upper portion and by the great profusion of individuals of species of Dielasma and Composita in which the brachidia are excellently preserved within the hollow interiors. The Miller limestone is characterized by an abundance of Aviculopectens, bryozoans, and productids.

Subzone C is best exposed on the Avon estuary at Windsor in the neighborhood of the bridges. The section, however, is broken by a number of important faults. The lower part of the subzone has frequent oölitic, algal, and sandy dolomitic beds that are sparingly fossiliferous, but in the upper part some 20 feet of platy blue limestone furnishes abundant Martinia and Productus.

Subzone D has a thick gypsum member at its base, in which are a number of thin calcareous seams crowded with ostracods and with the foraminifer Nodosinella. The upper beds are best exposed at the mouth of Kennetcook river, where 100 feet of thinly bedded blue-grey limestones carry a fauna which is more normally marine than any which preceded it. Cup corals become abundant here for the first time, there is a greater variety of brachiopods, and the molluscs are mainly confined to the Bellerophontidæ.

The following generalized lithological section illustrates the rhythmic recurrence of mud deposition with chemical and organic deposits:

D. Kennetcook limestone..

Gypsum and anhydrite with Nodosinella bands.
Red shale, etc.

100'+

Gypsum, anhydrite, and red shale.

B1. Maxner limestone

Gypsum, anhydrite, and red shale.

A. Basal limestone, conglomerate, and quartzite.

80'

Environmental factors. As interbedded chemical deposits and red argillaceous shales form so prominent a feature throughout the Windsor series, it is evident that the biotic conditions were decidedly abnormal, a fundamental fact that must be borne in mind in comparative studies with other faunas. The most significant factors of the Windsor seas were shallow waters, probable high temperatures, and varying salinities, culminating at intervals in conditions intolerable to a bottom life. These unfavorable conditions were not confined to local pools, but were of widespread extent in the Windsor basin.

The shallow-water conditions that prevailed were doubtless maintained by progressive subsidence in a manner analogous to the previous differential movements that controlled the terrestrial deposition in the basin. A "barrier" of some sort certainly existed against free communication with the outer sea, and it was more probably a tectonic upwarp due to the same differential movements than a depositional sand or gravel bar or a current barrier established by differential salinites or by controlling winds alone. The Windsor sea, moreover, was of a geosynclinal Paleozoic type and not a shelf sea in the nature of the present Rann of Cutch. There is, furthermore, no evidence at hand for complete isolation, and the "breaks" in the sequence are in the nature of diastems or minor disconformities. Stages of extreme shallowness took place in the middle of the epoch, as attested by dolomitic sands, algal bands, oölites, Modiola bands, etc., which beds are characterized by peculiarly restricted faunules, whilst the nearest approach to normal marine conditions was reached in the late life of the sea.

Desiccation rarely proceeded beyond the precipitation of gypsum or anhydrite, but this alone would demand, since there were no marked volume contractions, a surface inflow from the outer sea of four to nine times the volume of water in the basin under conditions in which the evaporation would approximately balance the accessions.

The times of gypsum or anhydrite deposition were not the only ones that prohibited the establishment of an

indigenous fauna, as red shales make up nearly one half of the total mass of the sediments, and so far as known are barren of fossils. In this case the extreme muddiness of the waters, combined with probable high temperatures (warm seas at this time are indicated by abundant reef corals in western Europe), was equally deleterious to animal life.

Faunal selection.-Unfavorable biotic conditions would affect various organisms differentially. They would be particularly prohibitory to the establishment of certain groups of greater sensibility, such as the corals, and it is not surprising that colonial corals have not been found in the Windsor deposits and that cup corals are very meagrely represented and gain only a temporary footing in the more normal marine waters of late Windsor time. An analysis of the Windsor fauna shows that of 104 species belonging to 63 genera, 48 species representing 22 genera are brachiopods, and 36 species representing 25 genera are molluscs. The fauna, therefore, as determined by the physical environment is essentially a molluscan-brachiopod one.

The composition of a faunal assemblage, however, is dependent not only on the direct environmental factor but on the available routes of migration. While the Windsor faunas show very little in common with the Mississippian faunas of interior America, there are clear affinities with the faunas that inhabited the seas of western Europe at this time. Accordingly, absentees in the Windsor faunas may likewise be absent or illy represented in the Avonian (Lower Carboniferous) faunas of the North Atlantic province with which direct migratory relations were established. This factor alone might account for the meagre representation of crinoids as mere stem fragments, for the absence of blastoids, and of such specialized bryozoans as Archimedes and Lyropora.

Correlation with the Viséen.—In correlating the Windsor faunas it is natural to turn in the first place to their nearest allies, the Lower Carboniferous faunas of western Europe. To the Belgian geologists, who were among the first to illustrate these faunas, we are indebted for the recognition of two major faunal divisions, the Tournaisien below and the Viséen above, corresponding approximately to the Waverlian and Tennesseean, respectively, of the American sequence. In recent years, British paleontologists, led and inspired by the admirable studies of