The argument this series makes, in 45 steps. It is the same spine every module is built on, and no module states more of it than its own share.
American share prices rose through the 1920s to a peak of about 381 on the Dow on 3 September 1929, and shares were being bought with a small payment down and the rest borrowed.
The Federal Reserve had raised interest rates in 1928 and 1929, before the crash, to stop people borrowing to buy shares, and tighter money reached the whole economy rather than one market.
The American economy stopped growing in August 1929, two months before the Wall Street crash, so the crash marked a fall that had already begun.
The selling came on two days that are still named, 24 and 29 October 1929, and by the low of 1932 shares had lost about nine tenths of their value - and the crash on its own is not what caused the Depression.
What followed was the longest fall ever recorded: forty-three months of shrinking output, from August 1929 to March 1933.
Work disappeared on a scale never seen before or since: about three American workers in a hundred were without a job in 1929, and about one in four by 1933.
About 9,000 American banks stopped trading between the crash of 1929 and the end of 1933, against a 1920s average of about 600 failures a year.
The failures spread from bank to bank and town to town, because a queue at one bank frightened the depositors at the next one.
Every bank that closed destroyed the deposits inside it, so the amount of money in America fell by roughly a third between 1929 and 1933, and prices fell by about as much.
That was about 12.8 million people out of work in 1933, and March that year was the worst month ever recorded.
The banking system stopped working altogether in March 1933, which is where the fall hit bottom: every state had closed its own banks before Roosevelt closed them all on 6 March.
A government promise stopped the runs: deposits were insured from 1934 up to $2,500 each, and nine insured banks failed that year against about 2,300 closures in 1931.
Britain went back to gold in 1925 at the same rate as before the First World War, which fixed the pound high and required British wages and prices to fall.
The rule was one-sided: a country losing gold had to cut back, and a country gaining gold was under no matching duty to expand, so falling prices were passed on and never offset.
Britain stopped swapping pounds for gold in September 1931, and sixteen countries devalued or left gold during that year.
Leaving gold worked because it removed the reason to keep cutting: a country no longer holding its money at a fixed gold price could let the amount of money grow again.
No country recovered while it stayed on gold, and the later a country left, the later its recovery came: the United States left in April 1933 and France not until 1936.
After 1933 the amount of money in America grew sharply, the monetary base rising by about half in four years.
Britain is the clearest case of the rule: it left gold in September 1931 and its recovery began in early 1932, before any other large economy's.
An early recovery was not a full one: unemployment among insured British workers peaked near a fifth and was still above 10% in 1937.
America raised import taxes in June 1930, and other countries put taxes on American goods in return.
Britain gave up free trade in 1932, putting a 10% tax on imported manufactures and preference on goods from the Empire.
World trade fell by about a third in volume between 1929 and 1932.
Economists are not persuaded that import taxes were a major cause, because the amounts of trade involved are much smaller than the tariff's reputation.
President Hoover did act - $1.5 billion of public works and a government lending body set up in January 1932 - and the collapse was far bigger than anything he tried.
The land failed at the same time as the economy: drought struck across the plains states through the 1930s, and farming families lost their living for reasons no policy in Washington had caused.
Pay and hours fell as well as jobs, so the people who kept work were often on lower pay and shorter hours.
Nothing measured any of this while it happened, because the United States had no national income statistics at all when the Depression began.
Governments built the statistics we now use because the 1930s showed them they had been governing without them.
What the state counted decided who it could see: the pensions law of 1935 left out farm workers and domestic servants, which is where most black workers were employed.
The damage did not lift: unemployment was still 14% in 1937, eight years after the fall began.
The unemployment figures now quoted for the 1930s were not collected at the time. They were worked out afterwards, in the 1940s, from other records.
The New Deal began with the banks: emergency legislation passed within days of 4 March 1933 and reopened the ones that were sound.
Two of its creations still govern American finance: insurance for bank deposits, and a regulator for the stock market.
Relief and jobs programmes put millions of people onto public payrolls.
The recovery law of 1933 let whole industries agree prices and wages between them, the Supreme Court struck it down, and some economists hold that it slowed the recovery.
A law of 1935 gave workers a legal right to join a union and to bargain with their employer through people they chose themselves.
The pensions and unemployment law of 1935 was passed by a government frightened of the movements outside it.
Recovery began in 1933, not in 1939 and not in 1941.
The recovery was incomplete: output came back and employment did not.
A second downturn ran from May 1937 to June 1938 and pushed unemployment back to 19%.
It happened because policy tightened too early: banks were made to hold twice as much in reserve, and taxes rose.
No nationwide programme of government spending on the scale of the gap was ever tried in the 1930s, and economists do not read the New Deal as one.
Full output and full employment returned during the Second World War.
Conscription and defence production took unemployment below the level it had been at in 1929: 1.9% in 1943 against 3.2% in 1929.