Thames Plate Glass Company

David Dungworth

Plate Glass manufacturing technology had been introduced to Britain from France in the later 18th century and for nearly 50 years one company (the British Plate Glass Company at Ravenhead, Lancashire) had dominated the industry. The early 19th century saw the establishment of a number of rivals, including the Thames Plate Glass Company in Poplar, London. The expansion in plate glass production was sustained by building booms and changing architectural fashions which prized large windows.

The production of plate glass was a complex and labour-intensive sector of the industry. Considerable attention was first devoted to obtaining the most pure ingredients possible. The sands used were the best available (i.e. with minimal iron content), with significant sources being the Isle of Wight and Kings Lynn. A variety of fluxes were used depending on technological, political or economic factors. By 1823 the Ravenhead works started to make use of soda from the double decomposition of common salt and potashes. Many also made use of Leblanc soda from the 1830s onwards. The source of lime used to stabilise the glass was usually chalk or high-quality limestone and this was often slaked prior to use.

The raw materials were melted in a crucible for 1 to 2 days and then transferred to a second refractory container (a cuvette) where the glass was further refined for several hours to remove any imperfections in the glass (especially bubbles). When the glass in the cuvette was judged to be ready it was lifted out of the furnace using a crane (Figure 1) and transported to a casting table which was positioned in front of an annealing chamber. The molten glass was poured onto the metal casting table and then flattened with a large metal roller, the thickness of the plate being determined by two metal strips placed along either side of the casting table (Figures 1 and 2). Cast plates of glass were generally half an inch (12.5mm) thick with an area (prior to cutting) of up to 100 square feet (10.8m2) in 1836, and up to 270 square feet (29m2) in 1876. The casting-tables were on wheels and typically weighed 14 tons (the plate glass formed on them typically weighing 0.3–1 tons). Once it had been rolled flat, the glass plate was pushed into the annealing chamber where it remained slowly cooling for several weeks. The tops of the casting tables were level with the floor of the annealing chamber to ease the transfer of the plate glass after casting (Figures 1 and 2).

Figure 1. Plate glass casting table showing the cuvette, heavy roller, and side rails (Knapp)

Figure 2. Casting plate glass (Porter 1832).: The illustrator appears to have reused (but not understood) the illustration in Diderot’s Encycopédié as this version is very similar but seems to show workers tilting the cuvette with their bare hands!

The annealed glass plates had rough surfaces that were not perfectly flat and required laborious grinding and polishing before they were ready for sale. The grinding and polishing was carried out by embedding a plate of glass in plaster of Paris to prevent its movement. The first stage consisted of grinding using sand and water which served to provide the surface of the glass with a plane surface. In some cases several different grades of sand were used progressively to grind the surface of the glass. Grinding was initially carried out by hand but the British Cast Plate Glass Company obtained a Boulton and Watt steam engine for grinding in 1789.

The next stage was called smoothing and made use of emery, a naturally occurring mineral largely made up of corundum (aluminium oxide) but often with various other iron-bearing minerals. Emery has long been valued as an abrasive because of its hardness and it was traditionally obtained from the Greek island of Naxos. Like the abrasive sand, the emery was usually used in several increasingly fine grades. The smoothing stage appears to have been carried out by hand even after the introduction of mechanical power for grinding. The final stage of the process (polishing) made use of carefully prepared iron oxide powder. This was traditionally obtained as crocus martis or colcothar, a waste product from the manufacture of sulphuric acid from iron sulphate. The grinding, smoothing and polishing would typically reduce the thickness of the glass by one half.

At each stage of grinding, smoothing and polishing of the cast plates they were carefully inspected for bubbles and other blemishes. Where these were detected, the plate was cut so that these now occupied marginal positions of little importance. This process, however, meant that relatively few cast plates survived the process intact and the price per square foot of plate glass increased with the size of the plate (Figure 3). The maximum size of plates of glass produced increased during the late 18th and early 19th century (Figure 4) with considerable improvement from the 1830s onwards.

Figure 3. Prices of plate glass (per square foot) in 1823

Figure 4. Maximum size of plate glass produced

In 1843 Timbs records that the ‘largest plate of glass yet cast has been finished by the Thames Plate Glass Company, at Poplar. Its dimensions are 14 ft. 8 in. long and 8½ ft. wide’. In 1845 Thames Plate Glass Company glass was selected for use in British light houses after tests carried out by Trinity House’s scientific advisor Michael Faraday. Faraday was scientific advisor from 1836 to 1865 to Trinity House (the body responsible for safe navigation round the shores of England and Wales).

In 1846 the Thames Plate Glass Company was mentioned in a paper read by Brayley to the Pharmaceutical Society on the subject of glass manufacture. Brayley explained the method of manufacturing cast plate glass with specific reference to the Thames Plate Glass Company. He lists the raw materials silica sand, lime, sodium carbonate, sodium nitrate and arsenic oxide. He further records that the sand was obtained from Kings Lynn and that the sodium carbonate was produced by the double decomposition of common salt and potashes, ‘in the way formerly pursued in the plate-glass works at Ravenhead’. He describes Thames Plate Glass Company glass as ‘remarkably free from colour’ and suggests that this was because the alkali used was completely free of sodium sulphate (soda prepared by the Leblanc process frequently contained some sulphur.

The Thames Plate Glass Company did not win the contract to provide the glass for the Crystal Palace (which was famously awarded to Chances of Birmingham) but at the Great Exhibition in 1851 it did exhibit the ‘the largest plate of glass in the world; its dimensions are eighteen feet eight inches by ten feet’. This plate was further described, ‘There is not a blemish on its brilliant surface, and it is as “true” as possible. It is placed in such a position that it reflects the whole length of the main avenue of the Crystal Palace, and the effect produced is superb’.

An account of a visit to the Thames Plate Glass Company was published by William Henry Wills in 1860. Wills was a successful journalist who contributed to and even edited popular journals of the period, such as Punch, the Monthly Magazine, Chamber’s Journal, the Daily News, and Household Words. The latter, which was a joint venture with Charles Dickens from 1850 to 1859, was the source of the articles which were published by Wills in 1860 as Old Leaves from Household Words. The article is written in a colourful and entertaining style.
Having, by this time, crossed a yard, we stood on the edge of a foul creek of the Thames, so horribly slimy that a crocodile, or an alligator, or any scaly monster of the Saurian period, seemed much more likely to be encountered in such a neighbourhood than the beautiful substance that makes our modern rooms so glittering and bright, our streets so dazzling, and our windows at once so radiant and so strong.
but many of the details appear to be consistent with other known facts, for example it refers to the recent introduction of mechanical smoothing which is mentioned by Muspratt writing in the same year.

During the early 1860s the largest plates of glass were in very high demand from some theatres. Henry Dircks had patented a method for producing the illusion of a ghost on stage (the technique uses the same basic principles as the autocue). A plate of glass was placed between the main stage and the audience which was used as a mirror to project an image from a second, hidden stage. The reflection produced by the glass plate was slightly translucent and so gave people or objects a ghostly appearance. Obviously the technique could only work if the plate glass had no detectable colour and was perfectly polished so that its presence was not obvious to the audience. By manipulating lighting on the second stage, the audience had the impression that people and objects on the main stage appeared and disappeared. The apparatus (essentially the plate glass and the second stage) was known variously as The Dircksian Phantasmagoria, after the inventor, or as Doctor Pepper’s Ghost, after Dr John Henry Pepper who with Dircks first displayed the apparatus at the Royal Polytechnic Institution, Regents Street, London in December 1862. The apparatus was a sensational success with an estimated quarter of a million people coming to the Polytechnic in the first fifteen months that the Ghost was displayed. In the year that followed, six theatres in Britain installed the necessary plate glass but the illusion required plates of 12 feet square which were the largest that could then be produced. Dircks further records that a visitor to the Thames Plate Glass Company in 1862 or 1863 was told that it had completely sold out of the largest plates that the company could make.

Despite the high regard in which the Thames Plate Glass Company was clearly held and the apparent high demand for plate glass, the company failed in 1874. The most likely reasons for the failure of the company are the actions of larger competitors, both domestic and overseas. From the middle of the 19th century Belgium, and then America, began to export window glass and later plate glass. The later 19th century saw the demise of many traditional plate glass manufacturers and the largest glass producers (e.g. Pilkingtons and Chances) bought out the smaller specialist plate glass firms and successfully moved into the plate glass business.

Recently the site of the Thames Plate Glass Company was excavated by Wessex Archaeology ahead of the planned construction of offices and flats. Later development on the site destroyed much of the archaeological evidence for the manufacture of glass, although traces of the annealing ovens and grinding rooms survived. Much of the archaeological stratigraphy consisted of a cream-pink mottled deposit which has proved to be a mixture of calcium sulphate and iron oxide. This presumably represents the use of plaster of Paris and hematite during the polishing process. Substantial quantities of plate glass were recovered showing varying degrees of finish from rolled but un-ground, through to smoothed (but not polished). The results of the scientific analysis of a variety of glass and associated materials will be presented at the AIHV meeting in Thessaloniki in September.

Return to Later and Peripheral Glasshouses.