eg in the above example you are summing B711712 - therefore 3 columns of data. So thinking about this operation in terms of access, why would you want to sum data from different fields. What dataa is stored in columns D,E and F. And why rows 711 and 712. What is significant about these.
This most likely indicates a normalisation problem in the access model.
What you SHOULD have is a query criterion that selects certain rows from your database, which correspond to the data in the excel rows 711 and 712 - and then some appropriate function that generates the total.
Easy when you get used to the Access paradigm, instead of Excel.
Not only queries, but breaking apart spreadsheets where one sheet holds combined data for several different things. Normalization is something like that, only more involved.