PCA results in weka explanation and extracting features from it

I’m working on KDD CUP’99 dateset. Training and test set are different datasets. after pre-processing I got 25 attributes. I applied 4 different classifiers through PCA in WEKA(classify tab). Now,I’m supposed to get a small subset of features with the best results in order to conclude that 1 of the 4 classifiers reduced the data set to n-features and got higher precision but I don’t know how to do that. how to specify these features while all I got is vector of attributes. I’m a little bit confused because it’s said that PCA is dimensionality-reduction technique not feature selection.

I hope I could explained my question clearly