Page 5 - gt1607
P. 5
Figure 1. Stages in the compilation of William Smith’s 1815 documented his observations; he would not have used it in the GSA TODAY | www.geosociety.org/gsatoday
Geology map by John Cary (showing part of Wiltshire). field because the scale was too small. Smith did make an early
experiment using a Cary one-inch to twenty-mile (~1:1,300,000)
supplied by a map that gives, in one view, the locality of map, but he said that the “maps were spoiled by speculating on the
thousands of specimens. By strong lines of colour, the ranges of stratification without sufficient data” (Smith, 1815,
principal ranges of strata are rendered conspicuous, and p. 27). His memoirs also record that he colored a map in the
naturally formed into classes, which may be seen and vicinity of Bath and also used the Day and Masters County Survey
understood at a distance from the map, without of Somerset (1782) (Phillips, 1844, p. 27).
distressing the eye to search for small characters. This is
the advantage of colours over any other mode of repre- In the early nineteenth century, scale adjustment was more
sentation. (Smith, 1815, p. 11) difficult but not insurmountable. In order to scale his field maps
The effect of Smith’s striking coloring technique also serves onto the 1815 base map, Smith would almost certainly have used a
to give added depth to the two-dimensional plane of the map: pantograph. Smith was very familiar with this instrument. In his
the individual strata appear to have a measure of three-dimen- journal entry for Friday, 18 December 1789, Smith records time
sional relief, which makes their relationship to one another spent “Repairing and Setting the pentograph [pantograph]”
more apparent. before he “began to reduce the Plan of a Mr. C. Norton’s allot-
ments” (OUMNH Archive, Diaries, WS/B0 p. 55). It is important
MAPPING SCALE to realize that although Smith’s final map may have been small
scale, its content was derived from his large-scale field maps.
The 1815 map measures more than eight feet by six feet and was
printed on 15 sheets. This map was the canvas upon which Smith PROJECTIONS
Smith would not have been overly concerned about the projec-
tion of his base map; however, projections are a crucial element of
any Geographic Information System (GIS). Projection informa-
tion does not appear on any Cary or Smith map. The graticules
around the margins of the maps show ticks of longitude and lati-
tude (usually based on the Greenwich meridian), but in most
cases there are no internal coordinate lines/ticks within the body
of the map. On the Cary Turnpike map, the latitude degree ticks
are angled to the frame, indicating that parallels may be curved
lines. Meridians may or may not be straight, but in all cases are
not parallel. However, although only a small-scale map, the
General Index Map for Cary’s 1794 New Map of England and
Wales and Part of Scotland does have internal longitude and lati-
tude lines. Some information is available regarding commonly
used late eighteenth-century and early nineteenth-century projec-
tions. John Snyder gives a good account of these projections
(Snyder, 1993); his list of projections was tested on the Smith and
Cary maps used in the GIS. First-order polynomial transforma-
tion (affine) was used to project map images to coordinates for
each of the tested projections. Residual (RMS) errors were calcu-
lated based on the framing graticules. Typical errors ranged from
>4,500 m (14,764 ft) to <1000 m (3281 ft). Some projections (e.g.,
Lambert Azimuthal Equal Area) achieved a good fit on the grati-
cule but had a considerably distorted shape. Apart from the
residual errors on the graticules, a judgment also had to be made
on the fit to places and coastlines. In this regard, it is worth
remembering that the maps are all about 200 years old.
Essentially, they predate the 1st Principal Triangulation of the
country, which started in 1783 and was not completed until 1853
(although the first phase was complete by 1796). Although the
general shape of England and Wales is excellent on all maps, it
does differ from the modern coastline. For example, the Cary and
Smith maps show the position of Lands End to be >4 miles north
of its actual position; similarly, the Isle of Man is offset by ~8 miles
to the northeast. Bonne, Cassini, and Transverse Mercator projec-
tions all achieved relatively good results. Bonne was marginally
the best fit on the Smith 1815 map, closely followed by Cassini
(used by the Ordnance Survey for the Old Series 1� maps of
5