GSGS 2823 1:250,000, War Office, August 1916, with ‘Secret’ red air information overprint

GSGS 2823 1:250,000, War Office, June 1918, declassified version

The 1916 map was produced during the Battle of the Somme (July–November 1916) for the Royal Flying Corps (RFC), and shows (in red) a staging airfield at Hawkinge in Kent, and the Front Line, Squadron Aerodromes and what appear to be Royal Flying Corps Brigade airfields/parks in France and Belgium. The most likely reason for its production was for the pilotage and navigation of ferry pilots and new squadrons when taking aircraft across the Channel, delivering them to base parks and aerodromes, and vice versa. There is also the possibility of a ‘strategical’ use in connection with the Battle of Flers-Courcelette (15 September 1916), an important stage of the great Somme battle, in which tanks were first used. 

What follows is an examination of the base (background) map and red 'Secret' overprint of GSGS 2823, August 1916, and a note on the June 1918 edition. This should be read in conjunction with the zoomable maps in a separate window from above.

An examination of the base map and red overprint of GSGS 2823 This map has two components:

1. The Background Topographical map, and
2. The ‘Secret’ red overprint showing air information

Firstly, the background topographical map at quarter-inch to the mile (England) and 1:250,000 (France & Belgium). It should be noted that this background map has no conventional sign panel in the margin, had no scale of yards and metres, has no marginal alpha-numeric reference system based on the 2-inch squares (this would be difficult, as the map carries the remnants of two conflicting sets of squares – for GSGS 2733 and GSGS 2738). For similar reasons, it does not (and can not) carry a geographic graticule. It has no magnetic rose or magnetic variation arrow, and no contours. In any case, contours were considered superfluous in an air map. It does, however, carry the spot height data in black which was on the black plates of the original component maps – in metres for France and Belgium, and feet for England (though there is no marginal information to this effect). All-in-all, it is a fudge!

 This topographical base has been compiled from three components:

1. The pre-war Ordnance Survey quarter-inch map of Great Britain.
2. The pre-war GSGS 2733 1:250,000 series of North West Europe (primarily Belgium), begun c.1911, based on the Belgian 1:100,000 series, and drawn on the Brussels meridian.
3. The wartime emergency GSGS 2738 1:250,000 series of France, based on the French 1:200,000 series, begun in August 1914 and drawn on the Paris meridian.

These three are examined in turn below: 

(I) The OS quarter-inch map of Great Britain The OS quarter-inch map of England and Wales used in GSGS 2823 (August 1916 edition) was part of sheet 20/24 of the current small sheet series second edition, complete with water-lining, and added colours: main roads filled sienna, woods green, and blue coastal waters. It can easily be distinguished from the first edition by the revised and added detail and updated names for railways (South Eastern & Chatham Railway rather than the former South Eastern Railway and the London, Chatham & Dover Railway which merged (operationally) in 1899). It was reproduced in GSGS 2823 at its original quarter-inch scale, rather than being enlarged to 1:250,000 to match the rest of the map. The coloured edition of the same quarter-inch map was used in some of the 1916 ‘Air Packets’ produced at the OS for the Admiralty War Staff’s Naval Intelligence Division.

 A completely new and updated drawing of (a larger area of) south-east England, at 1:250,000, was used for the June 1918 edition of GSGS 2823. This had previously appeared in the new edition of Sheet 1A of GSGS 2733/2738 (see below) which was printed in September 1917. In this new drawing, water-lining was discarded, names were larger and clearer to read, and railways and roads were similarly strengthened. Roads and woods were coloured, as in the 1916 edition. All this made the map much easier to read for airmen.

 (II and III) The GSGS 2733 and 2738 series These two series had quite different histories, and were initially cartographically distinct, but during the war they became more closely related, to the extent that there developed considerable confusion as to the series-identity of certain sheets.

• The pre-war GSGS 2733 1:250,000 series North West Europe (primarily Belgium), begun c.1911, was based on the Belgian 1:100,000 series (redrawn before the war on different sheetlines as the British 1:100,000 series GSGS 2364), and drawn on the Brussels meridian.
• The wartime emergency GSGS 2738 1:250,000 series France, based on the French 1:200,000 series (but not a direct reproduction), was drawn rapidly as a route map in simplified style at the Ordnance Survey in August 1914 on the Paris meridian.

Although the GSGS 2733 and 2738 series were drawn on the Brussels and Paris meridians respectively (hence the sheetlines at angles to each other), both carried a geographical graticule referred to Greenwich. 

Both series carried a 2-inch alpha-numeric reference squaring system, related to their meridian and sheetlines, and when sheets of the two series were juxtaposed it could clearly be seen that these were at angles to each other. This is apparent on the 1916 edition of GSGS 2823, where sheets of the two series were butted together without the 2-inch squaring grids being duffed out. 

An undated (but apparently 1918) index diagram, without a GSGS number, bearing the title Index to 1:250,000 Maps of France GSGS 2738, carries the following inscription: 

Index to 1:250,000 Maps of France GSGS 2738 

This series is a reproduction of the French 1/200,000 map, and in a similar style. Sheets 3a, 7, 7a, 13a and 23, are not published. East of the line A.B.C. [north-east corner] the area is also covered by the 1/250,000 map of Germany [GSGS 2740]. 

An identical index diagram and inscription was included in the Catalogue of Maps of the Theatres of War . . . Issued by the Geographical Section of the General Staff., Revised to 1st May 1918. This shared a plate jointly with an index map for GSGS 2733; the designation for the plate was GSGS 3281, OS 1918. This index diagram for GSGS 2738 showed sheet 1a (covering the Boulogne–Calais–Dunkerque area, to the west of Sheet 1 (GSGS 2733)), which included south-east England, as a large sheet, rather than earlier index maps which showed a smaller sheet. Incidentally, the sheet and index numbering throughout the war used 1a and 1A interchangeably. The 1914 edition of sheet 1a (square rather than rectangular) only showed a smaller part of SE England, in outline. 

Sheet 1A (GSGS 2738) was initially (1914) of smaller dimensions than the rest of the series. A simplified route map, derived from the French 1:200,000 like the rest of the GSGS 2738 series, and roughly 38cm square, it was ‘Prepared and Printed at the Ordnance Survey Office, Southampton, 1914.’ It also included a small part of south-east England in white, i.e. blank – showing no detail. The later edition of this sheet, produced at the Ordnance Survey in September 1917, was completely redrawn (in GSGS 2733 style and on 2733 sheetlines) and was larger (to take in more of south-east England as far west as London and Brighton) than the standard sheet dimensions for the series. South-east England was now shown in detail, redrawn as described above at true 1:250,000 scale.

 The 1914 (first) edition of Sheet 1A carried the marginal note:- As some sheets of this series have been prepared by the War Office and others by the Ordnance Survey, margins may not always exactly agree. A peculiarity was that the index map showed sheets 1, 2, 4 and 5 (all properly GSGS 2733, on the Brussels meridian and therefore with sheetlines at an angle to those of 2738, which was drawn on the Paris meridian) as being on the same sheetlines as 2738. It shared this feature with several wartime and post-war index maps, and suggested that certain of the northern 2733 and 2738 sheets had been redrawn and/or recast on each other’s sheetlines. In fact, two of the GSGS 2738 sheets, 1A and 3, were redrawn, in or after late 1916, in GSGS 2738 style, with more detail than previously, and on GSGS 2733 sheetlines, and became part of that series. Thus sheets 1A and 3 existed simultaneously in two different series – an interesting phenomenon! 

It should be mentioned here that, in 1915, as the British sector of the Western Front had been extended southwards, a new combined ‘Sheet 1 & Part of 4’ was produced, extending as far south as Douai (inclusive). This included revised detail, notably more minor roads. As we shall see, this sheet was used as part of the compilation of GSGS 2823, 1916.

 The 1918 GSGS 2733 index map mentioned above showed the six sheets, 1A, 1, 2, 3, 4 and 5, all on the same system of sheetlines, with no overlaps, confirming that 1A and 3 had been redrawn from their earlier 2738 forms and recast on 2733 sheetlines. This index diagram bore the note:- These 6 sheets cover almost the same area as the sheets of 1/250,000 France which bear the same number. They show much more detail. Study of various sheets, including administrative grey and blue combined sheets, shows this to have been the case.

 Sheets 1A and 3 (formerly of GSGS 2738) had been redrawn, both probably in 1917, with greater and more accurate detail and, fitted to sheets 1 and 4 (of GSGS 2733), and included in certain special administrative combined sheets. Sheet 1A of September 1917 was published separately under the designation GSGS 2738, and also under 2733. Possibly the same occurred with Sheet 3. On some printings of Sheet 1A the figure 8 of 2738 has been crudely altered to the last 3 of 2733. Confusingly, sheets of 2738 (France) were printed during the war under the cover title North West Europe (properly belonging to 2733). There was clearly a strong desire to unify the two series! For the compilation of GSGS 2823, WO August 1916, parts of five component sheets were used: ¼ inch Sheet 20/24 (2nd Edition) showing SE England; GSGS 2733, NW Europe, Sheet 1 & Part of 4, 1915; GSGS 2733, NW Europe, Sheet 4, 1914; GSGS 2738, France, Sheet 1A, 1914; GSGS 2738, France, Sheet 3, 1914.

 The right-hand (eastern) margin of GSGS 2738 is parallel to the north-south sheetlines of GSGS 2733, as can be made out from the 2-inch system of squares on the eastern part of the sheet. 

The drawing styles of the component sheets differed; in particular, the French part of Sheet 1A (drawn in 1914) carried names in a sloping sans-serif, in the same style as some of the sheets of Germany (GSGS 2740) drawn at the OS at this time. The parts derived from Sheets 1 and 4 had the serif roman of those (pre-war) drawings, while the part derived from Sheet 3 had sans serif names.

 Three early sources for information about the GSGS 2733/2738 series are the late-1914 edition (‘Reprinted, with additions, [November] 1914’) of the War Office’s 1912 Manual of Map Reading and Field Sketching, Colonel E. M. Jack’s late-1916 (printed in early 1917) General Staff publication Maps and Artillery Boards, and Jack’s 1920 Report on Survey on the Western Front. 

The 1914 edition of the Manual of Map Reading and Field Sketching stated: 

(5) The 1/250,000 map of North-west Europe [GSGS 2733, sheets 1, 2, 4 and 5], including Belgium and parts of France, Holland and Germany, is in the same general style as the 1/100,000 map of Belgium [GSGS 2364], but the symbols used for railways are different. The map requires but little explanation. Contours are at 50-metre, or about 164-feet interval, and heights are also given in metres. The rivers and their names are boldly depicted, which makes it easy to read the map quickly. [Useful for airmen!]. For the most part the country is so flat that the contours are not sufficiently close to give a clear indication of the shape of the ground. Sometimes there will be no contours between two rivers, and often there will be but one or two. Where there are none the country is practically flat, and where there is even one the contour height or value will indicate the extent of the rise.

 Most of the area covered by these four sheets is very thickly inhabited. Numerous scattered houses will be observed. Many more have of course been omitted, but the endeavour has been throughout to indicate the general character of the country.

 The Ardennes begin in the East of Sheet 4, and continue throughout Sheet 5, and these areas have a different character to the rest of the map. They are thickly covered with forest, and the contours are here close enough together to show the shape of the ground. The map is divided into 2-inch squares, which can be referred to by the letters and numbers in the margin. The length of the sides of the squares is 2 inches, or almost exactly 8 miles.

 (6) 1/250,000 map of France [GSGS 2738, eventually including sheets 1A, 2A, 3, 6 . . . 23]. Maps of other areas have been hurriedly prepared by copying the French map on the 1/200,000 scale and reducing it to the 1/250,000 scale. A large area of France has been prepared in this style and also two sheets [2a (Cologne) and 6 (Coblentz-Frankfort)] to the East of the area covered by the N.W. Europe series [GSGS 2733]. These maps require but little explanation, but it should be observed that the contours are at 40-metre intervals as on the map from which they were copied. [The new GSGS 2740 series of Germany also had contours at 40-metres]. It was also found necessary to alter the usual contour colour, and to print them in a subdued tint, in order to avoid as far as possible any confusion between the contours and the single-line red roads. The sheets of the N.W. Europe [GSGS 2733] and of the French series [GSGS 2738] are numbered consecutively, and are shown in one index [in fact not shown on any index diagram in this edition; the bound-in ‘Index to Maps of Belgium and North-East France’ (War Office, October 1914) only shows maps on the scales of 1/380,160, 1/100,000 and 1/80,000] on the sheets of the French maps. It will be seen from the index that there is a considerable overlap between the sheets of N.W. Europe [GSGS 2733, sheets 1, 2, 4 and 5] and the sheets to the East [2a and 6]. A new sheet to be called 2a is in preparation and lies immediately North of sheet 6.’ (Manual of Map Reading and Field Sketching, [November] 1914, pp.90-91.) 

Maps and Artillery Boards, written in France in 1916 by Jack at ‘Maps’ GHQ, had this to say about the GSGS 2733/2738 series: 

‘1:250,000, or 4 m[iles]. to 1 inch. – This map covers, in about 20 sheets, Belgium and Northern France down to Poitiers, Moulins, &c. The series is not uniform in style throughout. Sheets 1, 2, 4 and 5 [series GSGS 2733] were prepared before the war, and were published on its outbreak. In these sheets the usual form of British maps has been followed, i.e., roads have black outline and brown colour, and villages are drawn to shape. Relief is shown by contours at 50 metres vertical interval. On the first edition of these sheets many minor roads were omitted, owing to lack of time, but these have been added in later editions. A combined sheet of parts 1 and 4 is published. [i.e. Sheet 1 and Part of 4]. The remaining sheets [i.e. series GSGS 2738] were prepared and published after outbreak of war, and are [simplified] copies of the French 1/200,000 maps. Roads are shown in red [without black outline], and villages by a symbol (circle). Relief is shown by form lines at 40 m[etres]. V.I. [vertical interval]. These sheets were produced under great pressure and very rapidly. They are less elaborate that the earlier sheets and contain less information. They are also drawn on a different meridian [i.e. Paris instead of Brussels]. The 1/250,000 is a useful strategical or motoring map. [or flying map, it could have been added].’ (Maps and Artillery Boards, p.4.) 

Jack’s 1920 Report on Survey on the Western Front gives some additional information, and makes it clear that no 1:250,000 sheets of either series were available on the outbreak of war; sheets of GSGS 2733 were soon printed, but sheets of GSGS 2738 had to be drawn from scratch and plates made before printing could begin: ‘The 1/250,000 (4 miles to 1 inch) which covered Belgium in four sheets and extended from Dunkerque to the Rhine, had been prepared at the War Office before the war, and was completed shortly after its outbreak. It was printed in five colours (outline, water, woods, roads and contours), and was a first-class map of its kind. This map was extended [GSGS 2738] early in the war to cover the northern part of France down to about the 47th degree of latitude, the work being done at the Ordnance Survey. To economise time a simpler style was adopted for the new sheets, following to some extent the French 1/200,000 (e.g., roads were shown by a single red line). This series was based on a meridian [Paris] to the west of that adopted for the original four sheets [Brussels], so that sheet lines of the new and old sheets fell at a slight angle with each other. The new sheets were printed in five colours, and formed a useful addition to the cartography of the area, though, being produced at speed, they were less complete that the first four sheets.’ (Jack, Colonel E. M., Report on Survey on the Western Front, London: War Office, 1920, p.32.)

 Both the GSGS 2733 and GSGS 2738 1:250,000 series also saw service in the Second World War, as did the map which GSGS 2733 was intended to replace:- GSGS 2517 Belgium and the North East of France at 6 miles to the inch (2nd Edition). GSGS 2733, redesignated GSGS 4042, was revised from air photos, brought up-to-date for communications, and recast on new sheetlines. GSGS 2738 retained its old designation, and was also extensively revised from air photos and brought up-to-date. 

Secondly, The ‘Secret’ red overprint showing air information

 The ‘Secret’ red ‘air’ (as opposed to red roads on the base map) overprint comprises four features:

1. The front line running north-south
2. The 24 circles denoting Squadron Aerodromes
3. The rectangles, numbered and named: 1 St Omer, 2 Boisdinghem, 3 St André, 4 Fienvillers or Candas.
4. Hawkinge airfield near Folkestone in Kent (in this case the red quadrilateral patch was enclosed in a black circle).

NB: There could be more than one squadron at an aerodrome. On 1 July 1916 (first day of the Somme battle) there were 27 or 28 squadrons on the Western Front, but by 16 November (end of the battle) there were 36. 

The position of the front line dates the red overprint as the end of August/beginning of September 1916, which accords very well with the August date of the base map, which would naturally be prepared and printed first. The front line includes Longueval and Delville Wood (captured on 29-31 August), but not Ginchy (captured 9 September) in the British area, while it is ambiguous about Guillemont (captured 3 September). The conclusion must be that the map was part of the preparations for the great British attack of 15 September – the Battle of Flers-Courcelette – in which the newly invented tanks were used for the first time.

That leaves at least two questions to be answered: what is the meaning of the four numbered and named rectangles, and what was the purpose of the map? Adopting William of Occam’s principle of least complexity, we could assume that the rectangles represent the base fields of the four RFC Brigades on the Western Front, one with each Army (IV Brigade served Fourth and Fifth (Reserve) Army); from north to south these putative attributions are:

Given the fact that the South-East of England is featured, with Hawkinge singled out, it seems likely that at least one function of the map was to assist reinforcing squadrons to fly out from England to the Armies in France. (A 1918 edition of GSGS 2733/8 sheet 1A bears a stamp on the cover: Aeroplane Despatch, A.A.R., Lympne, which of course is quite close to Hawkinge). Ferry pilots could use the map to deliver machines across the Channel to the Brigade fields, and the map could also be used to fly machines forward to Squadron aerodromes. Given the proximity of the front line to the eastern edge, however, the map would not be very useful for operations into enemy territory, as the map hardly extends east beyond the line Ostend–Thourout–Lille–Douai.

 Other lines of enquiry about the meaning of the red rectangles have proved dead ends. The idea that they represented Aircraft Parks or Supply and Repair Depots was scotched by the fact that there were never more than two Standing Supply and Repair Depots – No. 1 at St Omer and No. 2 at Candas – which received all new aircraft from England or France (including a reserve), manufactured experimental fittings, and overhauled and reconstructed aeroplanes, balloons, transport and flying accessories, while the Parks drew supplies from the Depots for all Squadrons. However, each of the three existing Brigades formed a new Mobile Aircraft Park on 15 December 1915, these being at:

• No. 1 Army Aircraft Park First Army Aire
• No. 2 Army Aircraft Park Second Army Hazebrouck
• No. 3 Army Aircraft Park Third Army Beauval

These were joined by a new Aircraft Park, after Fourth Army was formed on the Somme front, on 1 April 1916:

• No. 4 Army Aircraft Park Fourth Army [Querrieu]

It will be noted that these Mobile Aircraft Parks were much closer to the front line than the four rectangles on the map. 

RFC Brigades, one for each Army, were created in a reorganisation of 30 January 1916, each Brigade comprising a Corps Wing and an Army Wing, while GHQ had the special 9th Wing for ‘strategical’ work. In April 1916, all fighting machines were concentrated in Army Wings, while the Corps Wings included the photography (for intelligence and mapping), artillery spotting (especially enemy-battery-location and observing fall-of-shell) and close reconnaissance. 

On 19 June 1916, the HQ of GHQ’s 9th Wing was moved to Fienvillers and Vert Galand in preparation for the Somme offensive. Its squadrons were responsible for strategic (long) reconnaissance for GHQ, distant bombing of communications, organised offensive against the German air service, dropping and picking up agents, etc. On 27 June, The RFC established an Advanced HQ in Fienvillers village, near the aerodrome. This was a 30-minute car drive from Haig’s Advanced GHQ at Beauquesne. On 30 March 1916, at the time that Haig moved his GHQ from St Omer to Montreuil, the RFC HQ Staff moved from their chateau at St Omer to one at St André-aux-Bois (south east of Montreuil). 

We must now consider another possibility. What the four red rectangles have in common is their proximity to RFC HQ and RFC Advanced HQ, and by association to Haig's GHQ and Advanced GHQ. They are all in the area to the rear (west) of the Squadron Aerodromes and Brigade Mobile Aircraft Parks:

1. St Omer } Both in St Omer area, near original RFC HQ
2. Boisdinghem } and GHQ at St Omer
3. St André [-aux-Bois] RFC HQ from March 1916; near Haig’s Montreuil GHQ
4. Fienvillers or Candas RFC Advanced HQ; near Haig’s Advanced GHQ at Beauquesne

All this suggests a possible vital, strategic function for the map. Any map showing British installations (e.g. airfields) would automatically be designated ‘Secret’, so we should not rush to the conclusion that there was some clandestine function, but there is a strong possibility that the map was to assist the rapid dissemination of ‘strategic’ intelligence, including air photos, acquired on 15 September over and beyond the Flers–Courcelette battlefield (including train movements and reserves moving towards the front) from the forward aerodromes back to RFC Advanced HQ and main HQ, and to Haig’s Advanced GHQ and main GHQ. In other words, it may have been a key component of the battlefield command, control and communications (3C) system, enabling intelligence and air photos to be flown back, and orders to be flown forward, supplementing the telephone and despatch rider letter service (DRLS). Wireless, while it existed, was unreliable and subject to interception, while coding/decoding, or ciphering/deciphering, was a lengthy and tedious business, unsuited to rapid response. 

Note on the June 1918 edition of GSGS 2823 

This edition involved a completely redrawn base map (i.e. it was compiled from the latest editions of the component sheets of GSGS 2733/2738). The whole map was now at 1:250,000. 

The north-western part, covering south-east England, used the completely new drawing at 1:250,000 which was first used in the September 1917 edition of GSGS 2738 (properly now 2733?) sheet 1A.

 The Franco-Belgian area used revised component sheets of GSGS 2733 (Sheet 1 and Part of 4, and Sheet 4), and the component sheets of 2738 (1A and 3) which had by then been revised and redrawn in 2733 style and on 2733 sheetlines.

 The whole Franco-Belgian area was now consistent in style. No contour plate or spot heights were included in the English part, but the Franco-Belgian area had height in metres on the black plate, as before.

No secret or other overprints have been identified.