Everything you are interested can be handled by the base or stats packages of R, which both came with your installation.

For complete lists of the functions you may be interested in try typing "?stats" and "?base" and then click to go to the index.

Also for graphics, e.g. histogram, this is found in the graphics package, (which is also included by default).

Also you may want to check out "An introduction to R" available at: http://cran.r-project.org/doc/manuals/R-intro.pdf

You can do awesome graphs with R. Most of the things which you mentioned can be handled by Basic R installation. If at all you need any specific statistical test, you can google and find the necessary package library which will available in plenty for sure. As Mathew mentioned try to read that manual which will give you the basic understanding of R.

Most of the things you want can be done with standard packages indeed. But you may consider to use RKWard (a graphical interface to R, http://rkward.sourceforge.net/). We implemented fairly decent summary plugins which give the described parameters and plots (boxplot, ...). The automatically generated code by RKWard is basically R code which can be used and modified as you require. The are some alternatives to RKWard mentioned in our paper http://www.jstatsoft.org/v49/i09/paper .

Thanks you all for the advice. My problem as stated is the large number of variables. I shoul also add that i am not cliver with programing in R. Certainly, basic stat package and functions will do. Thanks Stefan for your hint and your paper, RKWard is quite interesting, i do not have to program.

Thanks for you all.

You may be happy with r-commander, a graphical interface which allows to use menus instead of commands.

Thanks Kat. I had a look at RCommander which is well done, inparticular i can easily import data from Excel 2003/2007 with no worry Though i can not find explication for gried (inactive statistical tests like Kurskal-Wallis test. I have numeric variables and categories in my data and within the numeric variables some are not normaly distributed. I thought i could run K-W test. Only Wilcoxon test is active (black) in the Statistics menu. Any explication?.

RCommander contains only a selection of possible analyses methods. If you want to use the whole range, you should use the commands. On the CRAN-R website, there is a search function, where you should find several packages for the K-W test. A wonderful resource is, I find, the book "R in a nutshell". Hope you'll find what you are looking for!