1) You need quite a few significant figures in the expected gene frequencies, because the expected fraction of SS genotypes is so low.  Here, the total numbers are very large, so if you rounded the expected fractions to one decimal place, you would get a value of 0.0 for the expected fraction, and therefor an expected number of 0.0 for SS, as opposed to 254.98 when you take the calculation to four decimal places.  Careless rounding might lead one to believe that SS genotypes are strongly favoured, instead of strongly selected against as is actually the case!

The absolute minimum significant figures are displayed here for getting an approximate answer! Always use as much precision as possible in intermediate steps when using your calculator. All those memories and stores on your calculator are there so that you minimize *rounding errors.

2) After calculating expected numbers to a fractional amount, it is especially bad form to round down expected numbers to whole individuals! Expected numbers are often fractional.  For example, if I toss a coin once, the expected number of heads is 0.5, even though I will actually only get a whole number of heads (0 or 1).