Depending on the nature of the data (whether longitudinal, cross-sectional or both (i.e., panel), the number of observations could represent the number of firm-year observations from each country. This is to mean, in your case, in 2006, 8 firms (companies) were included in the sample and in 2004, there were 5 firms (companies) studied, making a total of 13 companies (firm-year observations). This applies to the other countries.
Belonging to your file, I see that the results are calculated for 2004 and 2006, it means that these authors calculated the cross - section data so that the number observation in this is the number of companies in each countries in year 2004 or 2006