When running a Monotype Composition Caster to cast from cellular matrices, the matrices are usually contained in a diecase, which holds a 225, 255, or 272 matrices in a grid arrangement. The caster positions this diecase to select the appropriate matrix over the opening of the mould where the type is cast.
These diecases use one of two styles of parts to allow the individual matrices a bit of vertical movement relative to each other, while still preventing the matrices from all falling out of the diecase altogether. In the diecases made in the USA by Lanston Monotype, the matrices are held by combs which wrap around three sides of a wasp-waist in each matrix. The combs either have 15 or 17 positions for matrices.
I was rummaging through my collection of Monotype diecase combs and found this mysterious set of combs:
This seems to be a near-complete set of combs for a diecase, including one blank bar (on the left), the two special combs that fit to the outside edges of the grid of matrices, and 13 regular combs that fit between rows of matrices. But the strange part is that these combs have positions for 18 matrices, and I am not aware of any diecase configuration that contains 18 columns of cellular matrices.
I don’t know if I also have an empty diecase around that would use these, and I wonder if these are leftovers from a now-disused experimental diecase design, just like the other odd diecase I have. Used as part of a 15×18 diecase layout, this would have made 270 matrices available for casting, but it would have required a new design for the rear airpin block, possibly using airpin combination F+N as another column selection (just as N+L and N+I combinations select columns “NI” and “NL” in the current incarnation of the caster). This larger diecase would have been made obsolete by the Unit Shift feature, which provided 272 matrices in the diecase and allowed more flexibility in character width selection (without Unit Shift, all 15, 17, (or 18?) matrices in each row must cast to the same width).
I should note that although the later Monomatic caster used an 18×18 diecase, this was divided into four 9×9 quadrants, each with its own set of 9-position combs, so these are not for the Monomatic.
Postscript: Mystery solved—as described in a newer post, the Monomatic had two diecase styles: the divided ones that I’m more familiar with, and undivided ones that would use exactly these combs.
I’ve solved the mystery and added a postscript linking to the answer.