Sir,

In spite of the considerable advances in our knowledge of drumlin formation resulting from field investigations and from physical studies of the basal flow of ice sheets, the origin of drumlins is not yet fully understood. Any new method of attack is therefore to be welcomed, such as the use of a computer on air photographs of a drumlin landscape (Reference VernonVernon, 1966, p. 401). Since the conclusions drawn from this exercise conflict, in some respects seriously, with those arrived at by field mapping before air photographs became available (Reference CharlesworthCharlesworth, 1939), it becomes necessary to examine the discrepancy.

The problem falls into two parts: first the origin of the drumlins, secondly and much more importantly, the source and flow direction of the ice that created them. In the field work for the 1939 map (Reference CharlesworthCharlesworth, 1939), based on the one-inch Ordnance Survey maps, all recognizable drumlins were mapped; their number in Vernon’s area was roughly 1470. Vernon chose “only those features with a definite elliptical shape”. His drumlins numbered 873. By so limiting the selection, Vernon of necessity failed to get a true average length/breadth ratio or density per unit area. The arbitrary selection also affected the spacing and the density pattern in bands both transverse and parallel to the ice flow and his conclusion that most drumlins “occur singly with no obvious relations to others” which is not true (Reference CharlesworthCharlesworth, 1939, map).

Vernon concludes that the drumlins are concentrated in linear groups across and parallel to the ice flow that possibly represent “waves in the ice forming belts of low pressure favourable to drumlin formation”. The conclusion is, however, weakened by the arbitrary selection and his implicit assumptions, viz. that the region is perfectly flat and that the rocks throughout are uniform in hardness, texture and composition. Both assumptions are contrary to known fact. Irregularities, whether inherited or ice-created, abound in the rock surface and the rocks vary constantly and rapidly over the 40 miles (64 km.) in the direction of the dip. Most significantly, the rocks strike roughly parallel with Vernon’s bands.

Clearly, while pressure waves may have occurred, no compelling evidence has been submitted that they existed in this area. An hypothesis which completely ignores two of the controlling factors in drumlin formation, the variations in the rock surface and in the composition and other characters of the local rocks, does so at its peril. A jury must bring in a Scottish verdict of not proven.

The directions of the drumlins on Vernon’s map and my map are in general agreement; contrary to Reference VernonVernon’s (1966, p. 405) statement, my drumlins in the area east of Bangor trend north–south as his do. The only discrepancies are first, in Vernon’s area VI where his trend is east–west and mine roughly west of north (in accord with the contours of the six-inch maps), and secondly in the area north of Newtownards where Vernon’s trend is north-west to south-east but on my map is north-south, in agreement with the independent witness of the Geological Survey Memoir (Reference LamplughLamplugh and others, 1904, p. 120) for both striae and drumlins in that area, and strangely enough with Vernon’s own direction of the striae which is also north–south (Reference VernonVernon, 1966, fig. 2).

Vernon concludes from the direction of the drumlins that the “picture is radically different from that given by Reference DwerryhouseDwerryhouse (1923) and Reference CharlesworthCharlesworth (1939), both of whom postulated a dominance of Scottish ice over north-east Ireland”. A glance at the two maps in question shows this statement to be a gross misrepresentation of two quite clear text-figures which represented the Irish ice as occupying the Lough Neagh basin and proceeding along the valleys of the Main and Lower Bann.

Vernon asserts that the striae, drumlins and erratics “all indicate a strong flow of Irish ice over north-east Ireland at least during the drumlin-forming stage of the late glacial period” and that the “Scottish ice, if present at all, was present only on the north of the Ards Peninsula”. However north-east Ireland be defined, it must surely include the eastern parts of the two north-east counties, Down and Antrim. Not one geologist in more than a century of research has ever found a single erratic in these two areas which was not either local or from Scotland. The Geological Survey Memoirs, including that of Belfast (Reference LamplughLamplugh and others, 1904) and the new edition of the Ballycastle area (Reference Wilson and RobbieWilson and Robbie, 1966) prove unmistakably that the Scottish ice alone has over-ridden these areas. This is in agreement with the direction of the drumlins in the Lagan Valley which “trend east-west or slightly south of west” (Reference LamplughLamplugh and others, 1904, p. 96), with the marine shells in the Belfast area (Reference LamplughLamplugh and others, 1904), with the evidence of the erratics (Reference CharlesworthCharlesworth, 1957, fig. 145), including the erratics of Ailsa Craig in the Dundonald area and Lias shells (from the east coast of County Antrim) at Castle Espie, both places just west of Vernon’s map, and with the equally conclusive evidence of the marginal drainage and glacier lakes which in the whole of this eastern area were created by invading Scottish ice (Reference CharlesworthCharles-worth, 1939, pl. XXVIII; Reference Charlesworth1957, p. 1212; Reference Charlesworth1963, p. 465, fig. 127). Thus “the ice stood across Belfast Lough north of Holywood [just north-west of Vernon’s map] and impounded a lake in the Holywood Glen which drained westwards” [my italics] (Reference LamplughLamplugh and others. 1904, p. 62).

In the quaking bog of the Irish Pleistocene, the invasion of East Antrim and East Down solely by Scottish ice is firm ground. This computer, by wrong feeding and interpretation, has led us into a realm of fantasy.

25 May 1967