Beware of Ordnance Survey Maps

[Joseph_Pestell]

Following up a thread on RMWeb (poster was looking for a small Caley terminus) led me to do a bit of research on Moffat and Beattock.

The National Library of Scotland site (so much better than old-maps) has, amongst many others, a 25" map from 1890s and a 6" inch map supposedly surveyed in 1938.

The 1890s map shows a small turntable by the locoshed and up goods sidings only extending as far as the first bridge north of the station. And the 1938 map (published in 1949) shows the same.

But other docs on the web include a signalling diagram from 1912 with a new turntable at the south end of the station and the sidings lengthened by taking them through the bridge.

I suspect the latter is right - drawn at the time of changes? - and the OS map wrong. How many other OS maps that we rely upon for our model researches are also this inaccurate?

[10800]

Read the small print on OS maps - 'updates' may only be partial, major roads for example, with many features being left unchanged.

[Tim V]

The recording of track etc features is not the primary purpose of an OS map. It is in fact incidental to the main purpose which is for rating. As such, do not rely on OS maps. They only give an indication.

[beast66606]

Difficult one - I wouldn't necessarily trust signalling diagrams either. The Lime Street team built a release road in one of their platforms and had to remove it again when it was pointed out that the road was proposed in an earlier plan and the work was not actually done.

[Guest stuartp]

They are generally very accurate in terms of the position of things, especially the larger scales, but not necessarily the longevity of things. They can only ever reflect what was there on the date they were surveyed, which might not be the date they were published. As has been mentioned, revisions often only included major changes and some map series were based on surveys and revisions of another series.

 

An example. The 1909 Six Inch Map shows a particular farm a few miles to the west of Harrogate. A photo dated a year later shows the same farm being demolished, a date supported by census returns and local authority records. The site was forested in 1915. However, the next edition published in 1956 still shows the farm, albeit now sitting in the middle of the forestry plantation. Every other detail of the 1956 map apart from the trees matches the 1909 map including the position of lettering and the typeface used, and it is clear that the 1956 map is merely the 1909 one with trees drawn on it. The plantation was a major alteration, the demolition of a single farm was not. The foundations of the farm are still there - once I'd cleared the mud and scrub off and surveyed what was left, the layout matched the OS's version almost exactly.

 

Some areas were revised several times, others once or not at all. Have a look at Bankfoot on the NLS site - you won't find the railway on the 25 Inch map (1:2500) at all as there's only one edition on there and the railway was built after it was surveyed.

[meil]

The 1938 edition was an emergency edition because of impending war. They were not going to be concerned with the niceties of railway sidings.

 

The map will have been based on the 192x edition.

[Peter Kazmierczak]

Remember too, that OS cartographers wouldn't necessarily know the intricacy of single and double slips, tandem turnouts, trap points, etc. Also detail beneath overall roofs and canopies would be omitted as well.

Edited February 5, 2013 by Western Sunset

[Joseph_Pestell]

The map in question says "Surveyed 1938" - without any qualifications or caveats. But as you all point out (love the story about the demolished farmhouse as we have a similar situation with a millhouse on our own farm), clearly some of these surveys are very partial.

 

So, as we all accept that we can not rely upon OS apart from the most general appreciation of the siting of the line (and sometimes not even that), where do we go for easily accessible info? Despite the thousands of railway books written, there are still many locations not covered in detail and most photos concentrate on locos which inconveniently hide the trackwork. 

[Orinoco]

Other inaccuracies include deliberate ones, like military installations. These were left as blanks on maps until the inter-war years - and sometimes much later.

 

It seems unlikely to me that we could ever get a totally accurate picture of what a station was like. Builders don't always follow drawings accurately. Even doing a survey of a current station today, there will usually be angles difficult or impossible to access, or something to obstruct a photo.

 

To make a model, we have to be prepared to make compromises and guesses to fill in the gaps. Striving for the perfect model may mean, in the end, that it just never gets built.

Edited February 6, 2013 by Orinoco

[Oldddudders]

The Ordnance Survey also seems to have had an "attitude" to its maps, which it deemed perfect. More than 40 years ago, I had a bachelor colleague who spent his long weekends surveying the countryside from his Commer van, comparing what he saw with maps. If he found errors of fact e.g. a radio transmitter on the wrong hill, he would amend the map. OS studiously ignored his amendments - failed to respond to his letters etc. Bartholemew, on the other hand, welcomed his updates, and I believe sent him new maps for his use and amendments.

[Joseph_Pestell]

The recording of track etc features is not the primary purpose of an OS map. It is in fact incidental to the main purpose which is for rating. As such, do not rely on OS maps. They only give an indication.

 

I have never thought about large-scale OS maps being primarily for rating purposes. But it explains why they meticulously list fields and their areas.

 

But in the Beattock case, the map would not serve this purpose as the new turntable was built on what was previously non-railway land occupied by the Craigielands blacksmith. Presumably, the District Valuer (probably called something different in Scotland) had a copy in their office that they drew amendments on.

[Guest stuartp]

So, as we all accept that we can not rely upon OS apart from the most general appreciation of the siting of the line (and sometimes not even that), where do we go for easily accessible info?

 

Just for the record, I don't accept that at all. Any 2D representation of a 3D landscape will have anomilies and quirks, and anything created by humans cannot be perfect let alone anything created by the government. They need care in their interpretation and coroboration against other sources never does any harm, but if you can find a more accurate record of what was there a hundred years ago which is as accessible and comprehensive as the OS 6" and 25" series then please share it.

Edited February 6, 2013 by stuartp

[IWCR]

Hello

 

All  mapping  may  have  some  errors/ommisions  but  the  OS  is  one  of  the  best  there  is.

Everything  shown  however  was  surveyed  by  one  person  and  drawn  in  by  a  different  one.  (2  possibities  of  error)

Many  updates  do  not  use  a  resurvey  at  all  but  information  only  for  specific  areas.

Details  of  a  new  housing  development  for  example  are sent  to  the  OS  by  the  local  council,  initial  updates  are  drawn  in  from  this,  (is  the  suppplied  information  right?)  later  this  may  be  covered   by  a  resurvey  perhaps  years  later  only  then  perhaps  finding  an  error.   Most  modern  resurveys  are  taken  from  sterioscopic  aerial  photos.

 

A  site  I  have  been  looking  at  for  a  potential  model  is  covered  by  a  25"  to  the  mile  OS  map,  I  also  have  the  railways  own  survey  at  40'  to  the  inch,  the  latter  is  a  much  larger  scale,  extremely  detailed  and  show s everything  very  well.  Comparing  the  two  shows  a  slight  distortion  in  the  overall  site  shape  hence  slight  variations  in  all  building  locations  etc.

The  site  is  on  a  considerable  slope  and  it  is  apparant  that  the  railway  survey  has  been  taken  from  measurments  along  the  ground  (ie  not  horizontal)  the  OS  survey  would  have  used  theodolites  etc  and  is  actually  the  more  accurate.

 

Pete

[The Stationmaster]

Interesting all this.  Several years ago I had a knock on the door which turned out to be someone carrying out update checks on behalf of the Ordnance Survey who was calling because we had a new house on the site.  She explained that they do try to get round to look out on the ground for changes and if any are reported they like to get details if possible - so I duly explained to her which new walls were where in relation to the drawn position of old walls etc and she made the necessary amendments.

 

So they do go round and check (whether they do it everywhere is something I don't know) although obviously they don't do it all the time.  And the call was clearly not for any other purpose - the Council Tax Assessment and Rating Officer had made her own separate calls and taken her own measurements (having said submitted plans were one thing, what subsequently appeared on the ground was quite often another and that she was only interested in what was actually there, and then only for rating purposes).  The latter was quite amusing because at the time we had demolished the old garage prior to building a new one - so the property was duly recorded as not having a garage, even though she was told a new one would be built and what it would contain etc.

[andyrush]

So, as we all accept that we can not rely upon OS apart from the most general appreciation of the siting of the line (and sometimes not even that), where do we go for easily accessible info? Despite the thousands of railway books written, there are still many locations not covered in detail and most photos concentrate on locos which inconveniently hide the trackwork. 

This is one of the reasons that I always reply to the 'Where can I find a track layout of...?' questions with references to RAILWAY plans, where I have details of these on our database. There are too many pitfalls to be able to rely on OS mapping absolutely, and the various textbooks on how to interpret them (Oliver etc) need reading as well.

[PhilJ W]

A lot of information on OS maps is or was supplied by local authorities. Some were better than others in supplying that information. Also a few years ago the OS found that another mapmaker was copying their maps illegally (without paying royalties) and they successfully prosecuted that company, which was Bartholemew.

[Tim V]

The map in question says "Surveyed 1938" - without any qualifications or caveats. But as you all point out (love the story about the demolished farmhouse as we have a similar situation with a millhouse on our own farm), clearly some of these surveys are very partial.

 

So, as we all accept that we can not rely upon OS apart from the most general appreciation of the siting of the line (and sometimes not even that), where do we go for easily accessible info? Despite the thousands of railway books written, there are still many locations not covered in detail and most photos concentrate on locos which inconveniently hide the trackwork. 

It is important to realise that, in research, there is no one source that will answer all your questions. The OS mapping is a source, but it has to be taken in conjunction with other sources like the railway's own survey (if done, if it survives), aerial photography (English Heritage have some for English sites), signalling plans, local authority plans, deposited plans for new railways (which can show previous installations). The fun is in tracking it down and coming to an educated decision. Over to you!

[DaveF]

A bit over 40 years ago Dad was building a model of of a closed station which wasn't too far from where he lived.  Some research was done from the remains of the station area, we then found that the various OS maps were contradictory, none of them were quite correct.

 

At the time I was working in my university vacations in the office of a big country estate and had access to the surveys they had made of the land they owned, which included land adjacent to the station, thoughtfully they had included the track plan.  

 

It always surprised me how different their own maps were from OS - and the estate were right, as I helped resurvey bits of their land and redrew the maps which ended up quite unlike the OS maps in places.

 

David

[Clive Mortimore]

When Collectors Corner was at Euston, I visited soon after a large number of maps relating to the changes to the Bedford to St Pancras electrification had been dumped there. Odd thing is many were OS maps not MR/LMS/BR (LMR) surveys. I acquired a few on that visit.

 

Another thing about OS maps not always being totally resurveyed is an out of date survey can often be to our (modellers) advantage. I purchased a copy of the 1977 edition of the 1/2500 map with Kings Cross station on it. The track plan was correct for steam to diesel change over, turntable in place and the single road diesel shed. That would make it 1961 to 1966? Just right for my dream model.

[buffalo]

...It always surprised me how different their own maps were from OS - and the estate were right, as I helped resurvey bits of their land and redrew the maps which ended up quite unlike the OS maps in places.

This is quite normal, though right and wrong are not simple issues in mapping. Surveying, like all measurement, involves errors and the treatment of errors is one of the more complex aspects of the discipline. Your survey was probably based on a short baseline within the estate and, perhaps a small number of additional fixed points were used to construct a network. This would have led to an internally consistent plan with relatively small errors. The network of points used by the OS was, of course, of a very much larger scale. Measurements from any point in a network include cumulative errors and those measurements taken in a larger network invariably have larger cumulative errors.

 

Because the OS would have been working over a very much larger area, they would have needed to adjust their results to minimise these errors. They needed to produce not only a reasonable approximation to the size and shape of a single estate, but to achieve the best approximation relative to everything else in a much larger surrounding area. Similarly, they need to make each set of survey data fit into the existing overal mapping.

 

Typically, the errors are lost (hidden) by adjustments in areas which are not of great importance to the surveyors and the majority of users of their maps. For example, you might find small areas of scrub or woodland that appear smaller or larger on maps than their true size, entirely because they have been chosen to adjust these cumulative errors.

 

You also have to remember that the earth is not flat and that all maps are projections from a curved surface onto a plane. Over the years, the OS have used several different projections in their published mapping, each of which has entirely predictable errors. For this reason, you cannot simply overlay maps made on different projections as they will not match up. Fortunately, these errors are relatively small and, if you need them, the OS and other mapping agencies publish formulae or tables needed for adjustment.

 

Nick

[sulzer27jd]

The NLS maps are a great facility and one that I use on a very regular basis. They are however only any good as a guide to the track plans. Cross-overs are regularly missed, points are shown in place of diamonds and other minor mapping errors are noted. The problem of course is that following the map would lead to significant errors on a layout. As is so often the case with research, use more than one source - especially a photograph.

 

J

Edited February 6, 2013 by sulzer27jd

[petethemole]

As I regularly work with old OS maps as part of my job I'm used to odd discrepancies; however most of my work uses historic 1/2500 maps within the City of Southampton that have been digitised for use with a GIS mapping system (Mapinfo). Direct comparison with the current map is easy and its surprising how accurate the old editions are, in general. As a major urban area it was re-surveyed several times between 1866 and 1933. The post-WW2 survey was at 1/1250 and done from aerial photographs taken for the purpose in 1946. I assume the whole country was done, although probably not all in the same year. The photos are held by the National Monument Record in Swindon (in the old Factory offices; the public access room is the Drawing Office). The basic tiles have been updated several times but erratically, some more frequently than others. Southampton as the headquarters of the OS probably got updated more than most places as the surveyors did it as training. Now it's all done from satellite images.

 

 The OS have been putting errors and spurious items into the maps for some time; it makes it easier to catch the copiers. 

 

Regarding the accuracy of the modern editions, I haven't found too much of a problem at 1/1250 or 1/2500 but when enlarged on the computer to (typically) 1/100 and compared with what's on the ground, there can be a lot of discrepancies. I worked on a major shopping mall development where the contractors GPS/Total Station survey was so different from the OS that our trial trenches were in different places on each set of maps.

 

For anybody interested in early railways, many large towns were surveyed at 1/500 and those town plans are an absolute gold mine of information, including track plans that are probably very accurate. Examples I'm familiar with are Southampton Terminus and Weymouth (in broad gauge). Local libraries and record offices should have copies of these. Southampton was also surveyed very accurately by a team of Royal Engineers in 1846. This map shows what must be the original layout of Southampton Terminus, with wagon turntables averywhere.

 

Pete

[Joseph_Pestell]

It is important to realise that, in research, there is no one source that will answer all your questions. The OS mapping is a source, but it has to be taken in conjunction with other sources like the railway's own survey (if done, if it survives), aerial photography (English Heritage have some for English sites), signalling plans, local authority plans, deposited plans for new railways (which can show previous installations). The fun is in tracking it down and coming to an educated decision. Over to you!

 

Copy one person's material = plagiarism. Copy several persons' material = research.

[Joseph_Pestell]

 The OS have been putting errors and spurious items into the maps for some time; it makes it easier to catch the copiers. 

Ian Allan used this excuse once to me when I complained about the number of errors in a railway atlas that they had recently published.

 

I replied that I thought 42 errors on just one page (including an EKLR station transposed onto the SER) was taking this a bit too far!

[Joseph_Pestell]

This is one of the reasons that I always reply to the 'Where can I find a track layout of...?' questions with references to RAILWAY plans, where I have details of these on our database.

Hi Andy,

 

Which database are you referring to, please? And how does one get access?