Three Rivers Glass — Glass Colors

Three Rivers made bottles and jars in two colours: flint — the glass trade’s term for colourless — and amber. That is what the period trade directory recorded, what the surviving marked ware shows, and what the collector record catalogues. This article gives the evidence, and then clears the three misreadings that recur when a piece is described by someone reading colour off a photograph.

The documented palette

The trade directory. The American Glass Review Glass Factory Yearbook recorded in 1934 that Three Rivers’ products “were made in flint and amber,” a listing amended in 1936 to “flint packers, fruit jars, beverage bottles, proprietary ware.” Across every yearbook entry from 1927 to 1936, no colour other than flint and amber is named, for the Texas plant or for the Bastrop, Louisiana branch [ev-0042].

The price list. Smith’s price list records colour only on the items where it mattered to a buyer, and only ever as clear or amber — the beer bottle in both, the Hoffmann-Hayman “Crystalvac” coffee jar in both, with the amber valued at roughly three times the clear [ev-0039].

The collectors. Two find-lists compiled by George West collectors, Hazel R. Gattis and Mrs. R. Y. Williams, are preserved in Thelma Lindholm’s marker research. Between them they catalogue dozens of marked pieces — fruit jars, pickle jars, Crystalvac jars, peanut-butter jars, milk bottles, beer bottles, pepper-sauce bottles — almost all as clear, with a couple of amber items: an amber Crystalvac jar and an amber beer bottle. No green, blue, or other colour appears in either list [ev-0041].

These three are genuinely independent of one another. A trade directory compiled from the company’s own returns, a collector’s price list built from his own examples, and two find-lists made by different people in the town where the factory stood are not copying each other, and they name the same two colours.

Amber was functional, not decorative

Amber glass is produced by adding carbon — coal, charcoal, or wood chips — plus sulphur, which reacts with iron in the melt. It blocks essentially all ultraviolet and blue light below about 450 nm, the wavelengths that drive the photochemical reaction that skunks beer and degrades other light-sensitive contents [ev-0009]. A general bottle house serving beverage bottlers therefore ran an amber tank alongside its flint tank, using flint wherever clarity mattered — soda and mineral water, food packers’ jars, medicine [ev-0009].

That is why the palette is two colours and not one, and why amber turns up on beer bottles specifically. It is also why amber examples are scarcer: the amber Crystalvac jar is the most valuable standard item in Smith’s catalogue [ev-0039].

Three misreadings

“Aqua” is not a Three Rivers colour. Aqua or aquamarine is the default un-decolorized tint of ordinary soda-lime glass, produced by iron impurities in the sand — what you get by not decolorizing, rather than a colour anyone chose [ev-0009]. Where the El Paso research states the colour of a Three Rivers-marked bottle, it is colourless: the Victory Bottling Works bottle’s first three variations “were all aqua,” but “the final variation … was colorless” and carried the 3 RIVERS {star} mark [ev-0048]; both documented Union Bottling Works Three Rivers bottles are likewise colourless [ev-0049]. A faintly aqua-tinted piece bearing the mark is under-decolorized flint, not a distinct product line.

The “green glass” line is about the sand, not the factory. A passage from Geology of Texas, quoted in the marker research, says the local Carrizo sand “fuses to produce clear, green glass suitable for the manufacture of bottles” [ev-0041]. That is a statement about what the raw material can do. The factory’s documented output is flint and amber [@ev-0042; @ev-0039], and the two statements are not in conflict — untreated soda-lime glass is green or aqua from the iron in the sand, and colourless glass is what you get after decolorizing it [ev-0009].

An “olive jar” is a shape, not a colour. The collector lists include an “olive jar” and, elsewhere, a “Clear, Small Olive Jar” — a jar for packing olives, recorded in clear glass [ev-0041].

Amethyst is not attested, and the chemistry is against it

No admitted source attests amethyst production. The chemistry makes it unlikely twice over.

Sun-coloured amethyst — colourless glass that purples under long ultraviolet exposure — comes from a manganese decolorizer, and manganese dominated from roughly the 1880s to the First World War. From about 1912–1915 through the 1950s the industry decolorizer was selenium dioxide, often with cobalt oxide [ev-0009]. Three Rivers started up in 1922 and ran to the late 1930s, entirely inside the selenium era, so its flint should not solarize to purple the way earlier glass does. And no source held here records deliberate amethyst as a production colour at all [@ev-0042; @ev-0039; @ev-0041].

A purple piece carrying the mark would therefore need explaining rather than cataloguing, and the explanation this expert cannot rule out is that the mark was misread.

Why the palette was narrow

The two-colour palette is what 1920s and 1930s bottle-house chemistry and economics predict. Colourless glass requires a decolorizer and not merely pure sand, because commercially iron-free sand was never practical [ev-0009]. The company’s advertised Carrizo sand purity therefore mattered in a specific way: purer sand means less decolorizer to reach true flint, which is a cost and quality edge, but the chemistry rather than the sand is what makes colourless glass possible [ev-0009].

Greens, blues, and cobalt served niche categories — poisons, cosmetics, branding — too small to justify a dedicated continuous tank, and glass tank furnaces ran continuously on one colour, so changing colour wasted batch material and fuel. The economically rational setup for a general packers’ and beverage house was two tanks, one flint and one amber, not a third [ev-0009].

What would change this

A genuinely green, blue, cobalt, or milk-glass piece with an unambiguous Three Rivers mark, photographed with the mark legible, would be a documented exception and none is on record. The weakest point in the account is that the trade-directory listings reach this expert through the Bottle Research Group’s reproduction of them [ev-0042] rather than from the yearbooks themselves; the yearbooks are not held here, and obtaining a run of them would let the listing be read directly instead of quoted.