I have been enjoying the comments on estimating wall declination, especially those which have the ring of experience about them.
I am prompted to offer the following Consumer Guide to the methods I have used over the years (and two I haven't). I am sure that many readers will be able to add to my list: 1. Architect's Drawings A typical Architect's site plan includes a very convincing ornamented letter N enfolding an arrow. The arrow usually points at some arbitrary angle with respect to the sides of the paper. Alas, the angle is also pretty arbitrary with respect to true north. I have known errors up to 15 degrees. Verdict: Avoid like the plague 2. Large-Scale Maps Good maps may well stem from very careful surveys but the representations of buildings are often more suggestive than accurate. You can often determine latitude and longitude to high precision but should treat the orientation of a depicted wall with great caution. You also have to worry (in the U.K.) about distinguishing grid north from true north. Verdict: Avoid 3. Magnetic compass Subject to numerous caveats you ought to be able to get good results. You have to be sure that there isn't a hidden iron drain or some such nearby and, of course, you need to know the local magnetic variation. The method is no use if the magnetic dip is too large. Verdict: Avoid 4. Using Solar Azimuth - General Carol Arnold, John Davies, Tony Moss and Bill Gottesman have described variants of the same idea: use the sun! I have used all these and more. I am not keen on plumb lines [they don't keep still and if you use a bucket of water to dampen the swings you just get a bowed string!]. If you make many observations widely spaced in hour-angle you can get a result better than 10 arc-minutes. Verdict: Fair to Good 5. Using Solar Azimuth - Window Ledge Method If you simply place a sheet of paper on a window ledge with one edge firmly against the inside of the bottom of the window frame you can often get the vertical edge of the window frame or a vertical glazing bar to cast a shadow. You draw a line just ahead of the (moving) shadow and note the time that the shadow reaches the line. Typical window ledges are close to horizontal and the sides of windows are approximately vertical so this can be quite accurate. Verdict: Fair 6. Using Solar Azimuth - Letting the client do it What do you do if you have to rely on someone else? The instructions I have given overseas clients (for example) are as follows: Find a flat board, a sheet of squared paper, a spirit level, a wooden pencil and a digital clock (preferably one that is radio-controlled and shows the date as well as the time). Place the board against the wall and place the squared paper on the board with one edge against the wall. Place the clock on the piece of paper. Check that the board is level and stand a pencil so that it balances on its blunt end (this is an added check that the board is level). Arrange that the pencil casts a shadow that falls across the squared paper. Take several digital photographs (preferably over several hours) which show the paper, the pencil, the shadow and the clock. E-mail me the results the same day. By counting squares I can get a reasonable estimate of the angle the shadow makes to the wall. The rest is as for 4 and 5. Make sure that you know how the time on the clock relates to UTC. Verdict: Fine if there is no other way 6. Using GPS An up-market approach using GPS kit exploits two GPS receivers each slaved to the other so that the relative phase-angles of the signals received from each satellite can be compared. You let the system run for 6 hours or so and via a good deal of software you can determine the position or one receiver relative to the other to about 5mm. Their absolute positions will not be known to such precision but that doesn't matter. Using a 50m baseline, a 5mm error means you can determine the azimuth of the baseline to about 20 arc-seconds. You then use standard surveying techniques to find the declination of your wall. You have to take care not to get stray reflections so you should be high up on something solid. Scaffolding will not do! This works splendidly even if it is cloudy. The snag is the cost. The kit costs about $80,000 and is obviously expensive to hire. It is suitable for high-budget sundials only! Verdict: Wonderful if you have a rich client 7. Cassini Method When setting out camera obscura noon marks a technique, which I think is due to Cassini, is first to note the point on the floor perpendicularly below the hole in the roof (immensely difficult to determine accurately) and then draw concentric circles round that point. Next you plot the hyperbolic path followed by the image of the sun on a given day and, with luck, the image will cross some of the circles twice, once before and once after noon. Taking these crossing points in pairs, find the mid-points. They should align with one another and with the point perpendicularly below the hole. This is your north-south line. Verdict: Fine if you are Cassini; I haven't tried it! 8. Using Stars A technique used by the surveyors responsible for digging railway tunnels in the 19th century was to follow a circum-polar star round the celestial pole and note its most easterly and most westerly points. Half-way between is due north. This sounds easy in theory but needs the right kit, the right kind of experience, a clear night, thermal clothing and considerable skill. I cannot imagine how they avoided freezing to death in the U.K. winters! Verdict: Not for beginners Incidentally, real walls can be a right pain! They aren't flat and they aren't vertical and you can easily come to grief. For large wall dials, a good deal of practical dialling amounts to a hard slog analysing survey data and undertaking laborious error analysis, but that's another story. Frank King Cambridge, U.K. -