Brett is native to many fruit skins - maybe not all fruit, and maybe not as populous as other strains, but still present. It would be fairly easy to isolate a single strain, you can just plate it and streak for singles. The trick to me is how to tell brett from sacch using things your average homebrewer has access to.
Growth rates? Flavor of finished beer?
How are you differentiating between S. bayanus/cerevisiae found on the fruit skins and "wild" yeast?
You're right, it's perfectly possible to have a "wild" variant of S. bayanus/cerevisiae which won't behave well when fermenting beer. I was thinking of other genera of yeasts that don't work well for any fermented beverage.
So where did a brewery get it's yeast back in 1850 or so? Did they just use wild yeast over and over again for a few hundred years until it mutated into something that made good beer?
Tom is basically correct, but by 1700, the large brewers had a pretty keen idea of how yeast worked and how to culture it, although they still didn't know precisely what it was. (That would have to wait until 1876 when Louis Pasteur published "Etudes sur la Bière.") In addition to the methods Tom has mentioned, breweries often borrowed yeast cultures from each other. House breweries often used bread yeast for beer and vice-versa, so homebrewers could get yeast from their local baker.
By the middle of the 19th century, the big breweries also had elaborate "clarifying" systems which could also be used to harvest yeast, such as Yorkshire Squares or Burton Unions. To some extent, yeast strains were influenced by this equipment.
Brewing using pure strains of yeast in brewing dates to 1883 when when Emil Hannsen figured out how to isolate individual yeast strains. Despite this, many brewers continued to use blends of different yeast strains (including Brett strains!) until the 1950s.
Way too much beer history here:
http://barclayperkins.blogspot.com/here:
http://zythophile.wordpress.com/and here:
http://www.europeanbeerguide.net/