The precision of your cell number depends on the concentration of the cells you count in your hemocytometer (there are machines that will do such counts, but initially, you will want to see what you are counting to make sure that the machine is counting the same thing you are). The larger the number of individual cells (you should not have any clumping) you count, the more accurate your estimate. Precision goes as N0.5/N, so for 100 cells counted, you have 10/100 = 10%. For 1000 cells counted you have ~ 3%, etc. You should count each of the nine squares (1 mm x 1 mm each) on your grid and load both grids on your slide to get 18 grids counted. Use a spreadsheet to get the average and standard deviation of the mean for the cell counts in the 9 squares in one grid and repeat for the other. Calculate the standard error of the mean for each of those 2 averages and if the means are not significantly different from each other, you can assume that both drops loaded into the slide are from the same population and recalculate the mean and standard error of the mean (sem) for the population in your solution (use all 18 well counts). Since the depth of the liquid in a hemocytometer is 0.1 mm, you can multiply the average count (amount in 0.1 mm3) by104 for the number of cells/cm3 = cells/mL (at 760 mm Hg pressure and 20 oC).
If you made you counting cell suspension from equal volumes of cell suspension and trypan blue (e.g., 0.1 mL aliquots of each), then you have to double your above estimated cell concentration to get the value appropriate for you original unstained cell suspension.
When dispensing your cells, based on the volumes needed for that number from your cell concentration measurements, if the number are low, the probability of the number of cells dispensed being what you planned follows a Poisson distribution, but if the numbers are larger, say 100s, then this will approximate a Normal (Gaussian) distribution.
If you are going to make ratios of colonies of cells that grew/viable cells seeded, i.e., Plating Efficiencies (PE) then you will need to use equipartition of variance to correctly account for the precisions (or imprecisions) associated with each count to get the correct precision estimate (sem) for the ratio (i.e., the PE).
MR. Miguel and Mr. Christopher are totally right, Mr. Miguel showed in few steps how to do, Mr. CHristopher showed the techinique with many details, I don't have contribuition to add.
Trypan Blue isn't as accurate for viability as erythrosin B dye exclusion.
Although steps escribed above for concentration can work, that is an extremely low cell density-- cells often need to work together to condition the medium, so are the cells going into only a microwell area?