Spreadsheet A has a tab for every month of the year. Each tab has a column
for meeting date and several columns for initials. The date column does not
have a date in every row, so out of 20 or so rows, only two or three cells in
the date column will be filled in. The same for the initials; i.e. just
because a date is filled in, there may not be corresponding initials.

I created a second spreadsheet with a single tab and linked the cells to
those in spreadsheet A. Because there is not a date in every row, my results
look like this:

A B C
Date Div1 Div2
1 01/15/08 CC 0
2 01/00/00 0 0
3 01/00/00 0 0
4 02/25/08 0 JM
5 03/03/08 MS MW
6 01/00/00 0 0
etc.

What I want: I need to use a rolling date range, so that the spreadsheet
looks at column A and pulls all dates between Today()-90 and Today() and
gives me a count of how many of those rows have initials filled in for column
B and how many of those rows have initials filled in for column C.

I have tried using DCOUNTA with the following formula:
=DCOUNTA(A6:B204,"CL",$A$2:$B$2), but I cannot get it to work.

I can get it to work using the COUNTIF function such as follows:
=COUNTIF(INDIRECT("'"&B$4&"'!A:A"),"="&$B$2)-COUNTIF(INDIRECT("'"&B$4&"'!A:A"),""&$B$2-90)

But this just gives me a count for the rows of dates.

Any help is appreciated.