time series --- seasonal adjustment I'm concerned to seasonal adjustment procedure and want to know the criteria for this purpose can anyone please give me the answer of the following question.
what should be the criteria for seasonal adjustment whether to do or not to do?
 A: Seasonal adjustment is the process of removing a nuisance periodic component. If you have a time series with small deviations in it, the true trend can be masked by the seasonal component. 
Therefore, it may be beneficial to remove this seasonal trend. A nice overview of this can be found in the matlab tutorial: http://www.mathworks.com/help/econ/seasonal-adjustment-1.html
Without any further information on what you are modeling it is impossible to tell you if seasonal effects will matter. For example, assume your time series has a fluctuation of 1000 units, now assume the seasonal effect only accounts for 0.01% of this. At that point it probably is not important to model and remove the season component.  On the other hand, if the seasonal effect accounts for 80% of the fluctuation then you do need to model and remove it. 
A: You are concerned that you have seasonality in you data. There exists a test-statistic called qs, that tests if there is any seasonality in your data or not. 
The statistic is defined as follows:
$$
QS = n(n+2)\left( \frac{R_s^2}{n-s} + \frac{R_{2s}^2}{n-2s} \right) \,,
$$
where $R_s$ and $R_{2s}$ denote autocorrelations of your series. 
