Sometimes on this thread people talk about MSAs, other times CSAs. Here is a link to a page at the Census Bureau's website, which explains the most recent criteria for determining both kinds of areas, as well as metropolitan divisions:
http://www.census.gov/population/www...s/00-32997.pdf. There is some material to slog through here, and the explanations of the criteria are several pages in, but the info is here for those who want to check it out.
Some of this info relates to questions raised on this thread. For example, the tread title poses a question about maximum distances between cities in the same metropolitan area. According to the Census Bureau's explanations of their criteria, for a county outside an MSA's core to be added to that MSA, there needs to be at least a 25-percent commuter exchange between that county and the MSA's core area, not just with the outer fringe of the MSA. You can't keep tacking new counties onto a metro area because of commuting to the outer fringe of the metro area. At least 25 percent of the employed population in the county has to work in the MSA's core. A county will also be included if at least 25 percent of its employment positions are held by residents of the core area who commute to work in that outlying county. This seems less likely than a commute inward to the core, but in either case, there is a practical limit on the distance between the farthest cities in a metro area, because there is only so far that a substantial number of people will commute. This maximum usual commuting distance is as far from an MSA's core as the suburbs will spread, which addresses the question in the thread title. It seems likely that West Palm-Miami is pushing close to those limits.
As for the question of if/when Boston's metro area will be "absorbed" into NYC's, you can be sure that the two MSA's will never merge, so long as metro areas are defined in any way similar to the current definitions. This is the case because of the practical limits on commuting distances. A drive from the core of Boston's MSA to the NYC metro core in the heavy traffic of commuting hours could take five hours one way. You're not going to see 25 percent of the workers in an area spend that long driving to work.
It's theoretically possible that the two CSAs could merge, but this is unlikely. On the pages linked to above, the explanation of CSAs includes the information that in addition to a commuting exchange of at least 15 percent between the metros there has to be a subjective perception by residents of the two metro areas of close ties between the metros in question. I have read elsewhere at the Census Bureau's website that in fact questionnaires about this issue were distributed with the '00 census forms, to people in areas being considered for addition to CSAs. This, for example, is part of the reason that the Providence MSA has recently been added to the Boston/Worcester/etc. CSA. According to questions the Census Bureau asked of local residents, there is a sense of close connection between Providence and these other metros, not surprising when one considers, for example, that there are commuter trains between Providence and Boston.
Clearly, the Census Bureau's inclusion of the local-perception requirement, in addition to a minimum degree of commuting exchange, indicates that they intend for the CSA concept to apply to an area that is broader than an MSA, but still fairly local. As KidYankee pointed out in post 23, there is a cultural difference between Boston and NYC (LOL about the description of CT as the "mutt," though I'm guessing that residents of eastern and northwestern CT, and the Hartford area, likely see their state as having its own identity). It seems likely that residents of the CSAs centering around NYC and Boston would view the two areas as distinct entities, with too little connection, and too much distance between the two areas' principal cities, to be considered one broad local region.