Boomer,
I looked at your photos again and scrolled sideways to get the whole picture. The bottom photo is from the front sight area, so the longer screw is showing less threads (3 for the front and 4 for the back). So the shorter screw has more threads engaged, was it designed that way or were parts swapped during a rebuild? Another possibility is the counterbore the screw head fits into is shallow, tolerance on the depth or a slight defect? The original drawings with tolerances for these parts are all gone. I have been machining since the early 70s and am used to modern tolerancing on drawings, I have no idea how tolerances were specified on these first guns with interchangeable parts. It was not too many years before the Krag that all parts were hand fitted.

On another note the sights that did have windage had screws with a lot longer head, so your front screw is not one of those.

Photos of all these sights with screws have been included on different places on KCA, maybe someone else can provide a link before Butlers Rangers gets back.

Parashooter wrote on Oct 1^st, 2018 at 3:42pm:

Thanks for the info!!!!   I've gone out and found/acquired a copy of that manual.

I am with you, it does not look right. Parashooter has been most helpful with the drawings and reality checks.
The rebuilds done by the arsenal resulted in a lot of parts being mixed on these old Krags. There might have been slight changes in the details on the screws and sights over time and parts were taken out of bins on the rebuilds and assembled with next part picked up. There is a real possibility these were just made that way, the screw head sticking out is not in a place were it would hang on things or cut an operators hands on a mangled slot.
The drawings used in the manuals were not done on CAD but by hand, so the proportions of the manual drawing and you your actual screws being different may mean nothing as the drawing is done by hand. I make drawings in CAD for CNC programing, then wipe out the numerical tags and add tolerances for actual running production. I will take the same drawing and remove the numbers again and rescale to metric for a technical drawing for a customer, all proportion in the drawing will stay the same in all the incarnations. Not so in the old days.
Another possibility is someone ran a couple of days production on a machine with too thick a head. Human nature being what it is we hate throwing stuff away. One or 2 parts discrepant while setting up a machine and I will toss them instead of messing with them. A couple of days worth of production and you ask if it will hurt anything if we use them. I read an account where a operator at Ruger cut a 1000 barrels and inch short. They became the basis for a short run of specials that collectors look for now.

I totally agree with Parashooter!

FWIW:
In my experience, most Krags with model 1892 and model 1896 rear-sights will have a 'thick-headed' screw anchoring the front hole of the sight base. I believe this dates back to the 1892 sight. It is OK.

There is a model 1896 front screw with a thinner head. (This may have been intended for carbine sight use).

The model 1901 sight had a totally different front screw (with no flange), this allowed it to anchor the lower-base, while allowing the upper-base to pivot.

The model 1898 and 1902 rear-sights had identical screws, front and rear.

The 'short' Krag sight-screws can often be interchanged and will work OK. This probably caused them to be mixed and interchanged over the years.
(This probably happened at Springfield, at arsenals, with 'field armorers', as well as, by hobbyists). 

It kind of goes back to the prototype. (I imagine the thicker screw heads might be a bit stronger and easier to make)?

Lets beat this dead horse again...

Poyer states that the 1896 carbine base was reduced to 0.520 inch at the front to conform to the ballistics of the 22-inch barrel.

I'm guessing the taller screw enabled it's use for both the 1896 rifle (which sat higher) and carbine (which sat lower) rear sights.

On the Springfield Armory prototype Krag, the rear of the sight-base appears to clamp a flange on the fitting attached to the front of the hand-guard. That may explain the need for a more robust rear-screw.

That tall rear screw-head on the prototype does not interfere with the 'hump' on the sight-leaf. But, if the leaf is put 'up' in the vertical position, pinching of a thumb might occur.

A flush rear screw-head makes sense on the form finally adopted. It would eliminate a potential dangerous edge.