It does need to be mentioned that this was a somewhat extreme solution, made by a railroad that could afford to 'do it right'. Let's look at this analytically:

The Big Boy had a high enough boiler pressure that the length of the gauge glass needed to be restricted. That meant that a solution like that used on NC&StL Dixie 576 -- a simple data plate indicating where proper 'water level' on a 2% grade would show -- couldn't be used on a Big Boy. (Theoretically you could have such plates on the two separate gauges for 'intermediate' ranges ... and some kind of damped clinometer next to them to show precisely what to 'inject for'...)

Some railroads used try cocks to back up a single water gauge. That is somewhat ... inadvisable ... on a 300psi engine. Therefore two gauges, one redundant, for normal running, and two gauges for the displaced range. (On an engine with a shorter boiler, you still might have four gauges with less vertical distance between groups, with perhaps the determining 'condition' being how long a glass you could effectively run safely for a given working pressure...)

On his presumably small, low-pressure engine I'd think a single longer gauge backed with try-cocks would be sufficient -- but I'll entertain all arguments to the contrary.

Keep posting, Robbie!

There are several pages of notes in the Southern Pacific shop manual regarding the leveling and setting of water glasses on several classes of locomotives dependent upon which route they were expected to operate on. On the larger locomotives stacked glasses was the norm for the Siskiyou route as well as others. It is a fun read. I recommend having a look.

Dear Dave

I am sorry I thought I would answer the question about crown sheet height from a water perspective instead of fuel. Yes I used a Big Boy because I had already done the calcs for someone else. I am sorry that I did not have the calcs done for an Allegheny as that discussion did not seem to upset you! I will try to do better in the future.

Robby Peartree

Not upset at all. Just wanted to bring it back into a context more relevant to the actual need of the inquiry.

Besides, I kind of liked thinking about what a 2 foot Big Boy would look like.

Well, since this thread is already derailed, here is a 14 1/8" Yellowstone for consideration. Currently at the Whiskey River awaiting to be regauged to 16" if I recall correctly.
https://www.discoverlivesteam.com/discoverforsale/forsale/8_SmithN/index.htm

If we're going to go down this path, what if the South Africans had chosen Mallets instead of Garratts? 2 foot Big Boys or something similar might have resulted.

Same with Nepal. Nepal had no choice because of the tight turns.

And they would have been operational disasters just about anywhere in Africa, not just South Africa or Rhodesia. Remember that at least one competitive trial between a Mallet design and a Garratt was actually conducted. The long fixed overhang complicated by having the whole rigid length of the firebox behind the rear-engine driving wheelbase makes a 'proper' simple articulated far less probable; a Challenger kludge with a long shallow firebox partly over that rear driving wheelbase will involve top-heaviness on Cape gauge at any comparable boiler capacity.

The only real 'alternative' to a Garratt would be a Meyer of some kind (with the deep firebox carried between the motor bogies, and scaled to tolerate the boiler being above the 'forward' engine driver diameter). That would be shorter than a Garratt, but probably involve additional bunker or cistern capability to provide adequate 'assured' weight on the rear engine, among a few other problems much easier to solve in a Garratt configuration.

This even before we get into vertical load and lateral thrust from a Mallet chassis of the size considered on typical African track construction that Garratts negotiate without incident. As I recall, even heavily-improved mains capable of relatively high speed still had very restricted axle loadings compared to American practice. The rear engine in particular might have substantial polar-moment issues at "Big Boy" proportional scale...

Bumping this to get it ahead of the now-dead 'certain threads'.

We can look more carefully at the waterside of a proper boiler. We have already noted the necessity of adjusting height to keep the crown adequately covered -- net of slosh -- on grades. Circulation is also a matter of interest.

Many of the original designs of Stephenson firebox treated the waterspace as thermal shielding, a way to water-cool the inner wrapper the way arc furnaces have to be cooled to keep from melting. The optimization of radiant uptake (which involves the fourth power of temperature!) is a major advance in 'super-power' designs, but it begins to involve prevention of 'departure from nucleate boiling' in taking up the radiant heat.

Most power boilers do this with a waterwall; in small highly-driven boilers like those suitable for locomotives, this begins early to benefit from forced circulation. A couple of the possibilities for circulation in the water legs include Lamont-style forced circulation (at a nominally high mass rate, up to 6x steam-generation mass flow, but at only slight differential pressure, a few psi, as the pumping is entirely within saturated water and steam at equilibrium, so low power and low pump maintenance.

A Cunningham circulator (patented in Canada in 1950) uses a jet pump, a bit like a modified injector, to move water from a 'downcoming' region of circulation in the convection section of the boiler to manifolds near the bottom of the water legs. This provides positive supply and circulation for the legs, more in a number of respects than 'natural circulation' largely driven by the effect of having the leg spaces widen toward the crown does.

Meanwhile, the Nicholson 'syphon' company made a great deal out of taking water from the bottom of the convection section (at the bottom throat) and using the 'widening' effect to provide both circulation and radiant uptake in flat-plate-sided, internally-stayed volumes in the combustion-gas plume. These were occasionally touted as helping to prevent crown failures when running with low water, the "idea" being that water accelerated up out of the syphons would wash across an otherwise-exposed crown and 'keep it cool'. In practice of course nothing of the kind would really happen, but you might get repeated random thermal cycling, not just of the overheated crown but of the transition zone between crown and syphons...

Note that a common thread that runs through these things is the relative absence of care in design to accommodate the water that is 'preferentially accelerated' through the legs by various enhanced-circulation ideas. Unlike the situation under the steam dome, there is no drawback to how high the "water" (it does not look at all like 'water', any more than fire in a firebox looks like something in a fireplace) can go, or how it flows once out of the legs. What is fairly clear to me, though, is that the enhanced volume ought to 'make its way' back to the convection section rather than just pile up over the crown with the two 'sides' and flow up the backhead colliding indeterminately somewhere up top.