While writing books and instructional materials over the years I’ve been forced to deal with this problem quite a bit , and I can’t say that I’ve been entirely consistent. I think Martin correctly identified the most difficult bit, as the other responses show. In some places, I’ve treated that whole stretch from the rise of the Manhattan to the death of the Old-Fashioned --say, 1880 to 1965–as essentially one period. In others, such as the B.A.R. manual, I’ve split it and taken the Dry Martini v Mai Tai 1960s as a new era.
These days, I prefer not to split it. That’s based on two things. One, a lot of time looking at drink menus, both American and not American, shows that the canon was very conservative throughout that long 80+ years, with cut-outs for Prohibition and the world wars, where things got simplified drastically. But at the end of each of those events, there were very strong efforts to bring things back to “normal.” Sure, a few new drinks came in in after each of those events, and sometimes new spirits came in (vodka, tequila) and old ones failed to come back (genever), but the heart of the canon remained remarkably resilient.
Tiki seems to me like a special case, because it was a parallel universe. But you can look through newspaper bar ads from the 1940s, 1950s and 1960s and go quite some time before finding a Tiki bar: it was never the dominant drink template and never replaced the canon.
In the late 1960s, however, the canon changed drastically. All kinds of new drinks and ingredients were in and plenty of old ones–rye whiskey, for example–were out. That’s far more of a change than what happened in the 1920s. I wouldn’t call this the Decline period, so much as the Alternative Canon or something like that. The Antithesis to the previous period’s Thesis (which means that, if you’re a good Hegelian, now we’ve got the Synthesis).
On the other hand, I find our closeness with the 20th century and the availability of sources leads us to exaggerate the effect of changes that aren’t all that significant when you step back a bit. The changes in Martin’s Formative Period were far greater, and yet we’ve got no trouble lumping the Julep made with rum, sugar, a sprig of mint and two small lumps of ice together with the one made with cognac, madeira and claret, with lemon, strawberries, segments of orange wheel, a mountain of shaved ice and a forest of mint. I don’t think we’re wrong to do that.