A monograph, The ACO Guide to 19th C. Sites (Ontario Ministry of Culture and
Recreation,  Historical Planning and Research Branch, London Office, N.d.),
has this to say about 19th c. window glass for Ontario sites.
"The best method of analysis is to measure with vernier calipers the sherd
thickness to the nearest tenth of a millimeter.  In sites dating to before
c. 1850 the average thickness will be below 1.5 mm whereas after this date
the average thickness will be greater than 1,5 mm (in fact, often about 2.0
mm).
 
A simpler, if less precise, method is to determine the percentage of glass
over and under 1/16" (1.55 mm) in thickness.  On sites earlier than c.1850
about 80% or more of the glass will be thinner than 1/16th";  on sites later
than this about 80% will be thicker than 1/16th". Obviously a site which has
both early and late 19th century occupations will fall somewhere between
these two figures.
A "quick-and-dirty" field test is illustrated below. [illustration omitted]
Randomly collect 10 (ten) sherds of window glass (these must be flat and not
twisted or distorted through burning) and then pile them into a stack.  A
site dating to the first half of the 19th century will have a stack which
measures about 1/2" (13 mm) or so in thickness;  a site dating to the last
half of the century will have a stack about 3/4" (19 mm) thick."
 
This does not offer help in distinguishing between 18th and 19th c. glass
and speaks only to Ontario assemblages, but the methodology might have some
transfer value.
 
Francis Carson
London Museum of Archaeology