A lower quadrant will as well, since the much heaver lenses and frames are on the opposite side of the pivot to counterweight the arm.
Before semaphore signals, the indications were given by railway policemen who held their arm out to indicate stop, but otherwise stand with his arm down by his side. Early semaphore signals mimicked that by having an arm that extended horizontally for "stop", but fell to the vertical for clear - in many examples it fitted into a slot in the post, although the dangers with that arrangement were made clear after the Abbots Ripton disaster, when several signals got stuck in their slots because of snow. (It would also give a "false clear" if the arm was missing altogether. Hence the movement to have the arm at 45 degress for clear, although other designs existed such as the GNR "somersault", in which the pivot point is bracketed out from the post so that the arm can be seen when in the vertical position. (Picture)
If signal wire broke, I agree the "slacked" weight of lens-frame will pivot the lower quadrant signal arm back up to "danger".
They also used flags - red for stop, green for 'go slow' - ie 'Caution' and white for 'All clear'
The original intention of the slotted post signals was to give the three degrees of warning - horizontal for 'stop'. 45deg for 'caution' and in the slot of the post for 'clear'.
The GN Somersault signal was a direct response to the Abbot Ripton's crash. They did retain the slotted post signals, but only used them for 'two-aspect' signalling.although other designs existed such as the GNR "somersault", in which the pivot point is bracketed out from the post so that the arm can be seen when in the vertical position. (Picture)
......
Freshly fallen snow has little affect on signals due to the rather small surface area that it can settle on. Remember, how much the arm can move is limited because it's connected by a steel wire to the lever in the signal box. And this wire will be under tension regardless of which position the arm is in (otherwise it would droop in-between the wire run supports and lie on the ballast or the ground).
This is often said but the crank is still remote from the arm. Any failure in the linkage from crank to arm could theoretically result in a wrong side failure. With upper quadrant, the final failure is the arm dropping off, which would be noticeable.
Something like an ice storm might cause blown snow to freeze onto a vertical surface such as the face or back of the arm, in which case the lower quadrant would be more likely to fail.
And the first upper quadrant signals were also three-position, introduced to avoid confusion between the "caution" on one of these and the "clear" on a normal signal. These fell out of use quite quickly, after which people realised the upper quadrant was generally better and most railways adopted it. The GWR, as per usual, was the exception.
The risk of the link between the semaphore signal arm and the weighted 'arm' or crank going wrong was much smaller than the risk of the wire run to the signal box becoming obstructed. The railway S&T (Signal & Telecom) Department was well aware of the risks, hence the signal mechanism as well as the wire runs were checked regularly.
Worth noting that in GWR and WR designs, the mechanical link up to the signal from the weighted arm at the base of the post where the wire is terminated, is a rod rather than a wire. One thing that helped preserve lower quadrants on the Western was a great reluctance to mix LQ and UQ arms at the same signalbox. Other railways and their successor regions seemed to have no compunction mixing them, at Shrewsbury and Banbury for example.
One accident where a (lower quadrant) signal may have frozen in the "off" position was Castlecary on 10th December 1937. The drivers of both trains concerned - a Dundee-Glasgow and an Edinburgh-Glasgow - were convinced that the Castlecary down distant was clear, when it should not have been. Ultimately signalman Sneddon at Castlecary was held to be mainly to blame for the very serious accident, although the driver of the Edinburgh train was also criticised for driving at excess speed in the snowy conditions.
Kensington Olympia in the late 70s/early 80s was a right mix.
I have heard of flagmen, positioned at signals in poor visibility using such mechanical aids to ascertain that the signal that they were guarding had been pulled to off. They could not see upwards in the dark and might require the shelter of their adjacent cabin in bad weather, making it difficult for them to hear the movement of the signal.
I'll draw upon the excellent insight in posts 3 and 5:
These definitely existed, known as fogsignalmen. They would be charged with putting detonators on the rails when the signal in question was at caution or danger, and removing them when it was cleared. Colour light signals were usually bright enough that they weren't needed, and they were no longer necessary once AWS provided an alternative means of alerting drivers to signals.
There's at least one accident report on Railways Archive where it is quoted that the fog was thick enough to make the signal itself not discernable from the bottom of its post. The lamp might be visible, but the beam through the spectacle plate is directed towards approaching trains and from below might be blinded out by any white light leaking directly downwards.
There are certainly reports of train crew, rather than signalmen, hanging an (empty) bucket on the balance arm at the bottom of the post in thick fog to give them an alert when the signal was cleared. It was particularly used by freight train crews held in a goods loop where they were some distance from the signal box, I have read.
Thanks John - that sounds a far more practical (and easier and no doubt more effective) approach.
Only in very specific circumstances...
I suggested it was a fogsignalman. As well as the points you make, an actual signalman would be controlling the signal, so have no need of a bucket to know what the it was showing.
