I’m trying to do a tech history piece on Western Union. Not trying very hard at the moment, mainly focusing on courseware. I wondered how big WU’s research department was at the height of telegraph dominance in the 1920s. The answer: Not very impressive. But they used their relatively small staff effectively, covering a wide range of topics from termites in poles to microwave circuitry.
A 1927 list of all industrial research labs includes many non-surprises and a few surprises. The book didn’t cover government or academic research, so it didn’t catch the Signal Corps work on radio or the agricultural and engineering research at land-grant schools. The intro also said the survey didn’t check or filter the responses. Some were dubious, claiming one employee who occasionally did research.
First the geography. Where were the research labs?
The Excel file is here. The biggies were unsurprising. Mass, NJ, NY were the research hub until Share Value killed industry. The whole rust belt was also big. The MISSING states were more interesting than the big ones.
Oregon and Washington were hardly on the map. Comparing each with the nearest alphabetical state:
Oklahoma 6
Oregon 3
Washington 7
West Virginia 11
The Oregon listings seem to be small branch offices of national companies, more like sales reps than labs. The Washington companies were in bricks, cement and fisheries. Boeing was already large at that time but didn’t list any research.
= = = = =
Research by category was dramatically different 100 years ago. Most of the activity was in materials and chemicals of various types, not in devices or machines. I didn’t try to count these, just separated the Big from the Huge.
Adhesives
Automobiles
Beverages
Building materials
Waste byproducts
Ceramics
Chemicals (huge)
Dyes
Electrical
Electrochemistry
Fats
Foods (huge)
Foundry equipment
Fuels (huge)
Insulation
Iron and steel
Lubricants
Metals and metallurgy (huge)
Paint (huge)
Petroleum
Plastics
Paper
Rubber
Textiles
Water and sewage
Wood
What’s partly missing, other than the obvious computer stuff? Radio was small. Office equipment was small. Pharma was small.
= = = = =
The biggest companies were also unsurprising. You could really condense this entire blogpost into one name:
Bell Labs.
Dupont, GE and Standard Oil were split up into a dozen different subcompanies and divisions, so I’ve probably underestimated them.
Bell Labs 2000
Dupont 1000
GE 700
Westinghouse 400
Standard Oil 300
Union Carbide 200
Dow 200
Kodak 200
RCA 100
Goodrich tires 100
Atlantic oil 100
Newport dyes 100
Raytheon 50
Western Union 50
NCR 50
= = = = =
In automobiles:
International Harvester 700!
Seems unlikely but the numbers are pretty clear.
GM 300
Packard 30
Studie 30
REO 30
Chrysler 20
White trucks 20
I didn’t see any other auto companies. Most had design staffs in the same 20-30 size range. Perhaps they didn’t respond to the survey or weren’t asked.
= = = = =
Who was known to be big but NOT on this list?
Ford! I know that Ford had as many product and basic researchers as GM, so it must have been unresponsive to the survey.
IBM or its ancestor CTR.
Polistra's Mill 