164 P.G.H. Boswell— Quantitative Methods in Stratigraphy. 
source of the Pliocene material was from the area of the Ardennes, 
etc., on the south-east, but that the minerals of the Kocene beds 
were derived from an entirely different direction, possibly from the 
west or south-west. The overlying Lower Glacial deposits were 
: derived, as their boulders also indicate, from the north, and possibly 
the north-east and north-west. These are broad generalizations only, 
but the subject is susceptible to more exact treatment which would 
lead to a correspondingly closer realization of ancient changes in 
geography. 
Unconformities are usually emphasized by the changes in mineral 
composition, and these support paleontological and field-evidence. 
Changes in the distribution of land and water and in the direction 
of the large rivers are thus revealed. As an outstanding example 
may be mentioned the contrast between the fine-grained residues, 
consisting largely of staurolite, kyanite, tourmaline, hornblende, and 
pyroxene, which are characteristic of the London Clay in Suffolk, 
and the coarse muscovite, red garnet, andalusite, staurolite, epidote, 
etc., of the overlying Boxstones at the base of the Crag of Rupelian, 
Miocene, or Diestian age. 
Recent work has possibly had a tendency to lead geologists to 
expect too much in certain directions from the comparative study of 
the residues of sediments, and it should therefore be stated at once 
that it is improbable that mineral constitution will have any cor- 
relative value over wide areas, certainly not comparable with that of 
fossils. From the nature of the subject we should expect various 
portions of basins of deposition to derive their material from different 
directions and sources. The writer has, however, endeavoured to 
show recently that mineral constitution, when fossils are rare or 
lacking (and if present, of wide range), has a distinct stratigraphical 
value over limited areas, when abundant samples are collected from 
numerous localities and horizons.t For this purpose the mineral 
assemblages of all the divisions of each of the geological stages of 
a district must be known. In connexion with these statements it 
should be said that the mineral assemblage of the Lower Greensand 
over most of its outcrop very closely resembles that of the Reading 
. Beds of South-East Suffolk and Northern Essex, which beds in turn 
are extremely similar in mineral composition to the Bagshot Beds of 
the area around Claygate and Oxshott in Surrey. Again, the mineral 
constitution of the various divisions of the Eocene beds in Kast and 
West Kent is much more similar throughout than in the corresponding 
divisions in East Anglia, where there is a greater variety of minerals. 
The respective Thanet Beds, Woolwich and Reading Beds, etc., in the 
two areas do not resemble each other in petrology, and no correlation 
could be attempted on such evidence alone. The composition of the 
Thanet Beds around Lille, as detailed by L. Cayeux,? is again 
different from either. ; 
Nevertheless, there are broad groupings which hold over a con- 
siderable area. Although the members of the Eocene Series differ 
among themselves, and each member may vary in composition over 
1 Abstr. Proc. Geol. Soc., No. 973, p. 76, etc., 1915. 
2 Ann. Soc. Géol. Nord, vol. xix, p. 264, 1891; also p. 90. 
