What would a small belt-driven shop have been cranking out around 1905?

My money is on a general jobbing shop, meaning they made repair parts and short runs for whoever walked in the door. The giveaway is the mix you're describing. If a shop was set up to make one product, like sewing machine heads or bicycle hubs, you'd see one or two part types in serious quantity, probaly with some filing or drilling jigs and fixture pieces in the crates too. Bushings in every size, stepped shafts in batches, and assorted knobs is exactly what a six-machine shop cranked out when the local mill, the cannery, or a farmer with a broken line shaft bearing showed up with a broken part and no drawing.

The blank parts actually support this. Catalog parts got stamped with the maker's name. Contract work for a mill or an equipment company often went out unmarked on purpose, since the customer didn't want their supplier undercut. And 'No. 4 fitting' in pencil sounds like a job number or an internal part number off a work order, which is how those shops tracked a run of forty bushings for one customer. Look at the crates again and see if any have multiple labels or old shipping marks underneath. Jobbers reused crates constantly. The knurled knobs might be the one real clue. Small instrument-style knobs showed up on everything from surveying tools to textile machinery controls to early electrical switchgear. If you've got a textile mill history in your town, I'd lean that way, since mills ate small brass and steel parts constantly and were ussually the biggest customer within a buggy ride. Two things worth trying. First, pull city directories from 1905 to 1915 for your area and find the shop listing. Those directories usually say what the business did, like 'machine shop' versus 'sewing machine needles' or whatever. Second, check with the local historical society or library for a whole册... a whole run of the local paper. A shop that size probably placed the same small ad or one-line notice for decades. If you find the name, patterns like what they advertised will settle it fast.

Those unmarked brass bushings would've been stock for the shop itself, half the old-timers kept bins of 'em in standard sizes and machined to fit on the spot. My grandfather ran a two-lathe place in the 40s and his bins looked exactly like what you're describing. Also don't skip the ceiling hardware, if there's a countershaft or clutch pulley up there, that tells you what speed the machines ran and hints at what they were cutting. And check the crate corners, pencil numbers often get written twice, once on the slat and once inside the lid.

Measure the knob threads before you guess any further. If they're oddball sizes like 27 threads per inch or some metric oddity, that points to instrument or electrical work, since those guys loved nonstandard threads to lock out the aftermarket. I restore old bench instruments as a side thing and about a third of the knobs I've run across are some thread you can't buy a die for. Also check wheather the knob bores are split with a little clamping screw. Split clamping knobs almost always mean something that got adjusted by hand often, like a focus or feed control, not a plain set screw mount on a mill lever.

One thing I'd add before settling on the jobbing theory, look at the shaft batches closely and see whether the steps follow the same ratio across the whole tray. When a shop was feeding one product line, even an unmarked one, the proportions repeat because they were working off the same hand-sketched pattern sheet for years. I went through a box of turn-of-the-century shafts from a millwright's estate and the family resemblance was obvious once I laid them side by side. Also worth flipping that pencil label over and viewing it at a low angle with a flashlight, older writing underneath sometimes shows as an impression in the wood even when the surface mark has faded.

Look down at the floor, not up at the ceiling. The bolt holes and pads will tell you what was in there. A ring of small pads means light machines and walk-in repairs. One heavy pad off by itself means they owned something bigger, and a shop that sprung for one was usually feeding a single product, not jobbing.

The foreman I apprenticed under came up in a belt shop around that era. They kept a day book, every job written in by hand, and he'd chalk the job number right on the crate. That pencil label smells like a day book entry to me. If any paperwork survived in the family, that's the thing to hunt for.

Update: the thread suggestion paid off. Measured the knobs and they're some oddball fine pitch, nothing standard, which fits the instrument theory. But the shaft steps are all over the place, no repeating proportions I can spot, so that crate label still has me stuck. Is 'No. 4' the kind of numbering a jobbing shop would use for repeat orders, or something else?