Linda Hannas divided puzzles into three groups in order to get a rough picture of early puzzle production. The groupings were based on who might choose the puzzle for a child: a governess, parents, or favourite aunts. A favorite aunt would only present you with the kind of puzzle that was just for fun.
Her sampling of English puzzles showed that as a percentage of the total, map puzzles declined slowly and linearly over time. Religious or strictly educational puzzles dropped sharply, by mid 19th century they made up less than ten percent of production. Many puzzles were still meant to be educational, but the intent was balanced with the idea that learning could be fun. This type of puzzle would fall into the parental group. As puzzle production entered the 20th century, the majority of puzzles were produced for children, meant to be fun, and still pretty much were being produced by manual processes.
Many of these puzzles were produced for years, some for decades. It is sometimes difficult to know or even guess at an exact date of when a puzzle was produced. It is quite reasonable to assume that some of my examples were manufactured in the first quarter of the 20th century. If one of these puzzles is in a wooden box with a copyright date, I am confident that the date is a good indicator. If the puzzle is in a cardboard box, even with a 19th century date, my first instinct would be to guess that it was from a 20th century production run and quite possibly a re-production made later in the century. The organization of these puzzles by date is my attempt to show you when these puzzles may have begun production, and not necessarily the true date of my copy of the puzzle.
What I have to display here are American-made puzzles, the earliest from about the mid 1850s, and most (but not all) fall into the "favorite aunt" category.