generate a distribution

hello world !

context: I have about 1e22 samples that have an average size of third-9. I could content myself with this information to characterize this population but it is not enough. Indeed, the size of third-9 is a medium in reality there are samples that have a size of eight fourth-and other first-12.

solution: I know that the size distribution of my sample follows a log normal reversed. So I can, knowing the average value, the number of sample mean and standard deviation in software such as MATLAB to generate me a size distribution that meets these criteria.

problem: I can not understand how I can do this with software. - I decided to MATLAB because it is a very powerful software that is able to do this in one line.

The trouble is that I expected to do this: [size, number] = logninv (1e22, 3rd-9, 0.05)

and have direct variables in size and number distribution desired. and drawing number according to size to get my distribution like a histogram : <http://www.google.fr/imgres?q=histogram&um=1&hl=fr&sa=N&biw=1600&bih=735&tbm=isch&tbnid=DQVF5Oamf1nQHM:&imgrefurl=http://www.vistadatavision.com/reports/histogram/&docid=4932O0wS-rWiQM&imgurl=http://www.vistadatavision.com/uploads/images/reports/histogram_rh_48.png&w=650&h=457&ei=fAj_TqWVBcOdOpvG6JgB&zoom=1&iact=hc&vpx=478&vpy=286&dur=1179&hovh=188&hovw=268&tx=190&ty=121&sig=104039641612689214034&page=1&tbnh=121&tbnw=172&start=0&ndsp=33&ved=1t:429,r:10,s:0>

[Image Analyst]

"I have about 1e22 samples" - Are you SURE you have 1e22 samples? That seems like such a huge number. And why is that "not enough"? How did you collect so many samples? If you collected 1 million samples per second it would take you 3.17 billion years to collect 10^22 samples. Also, I don't know what "third-9," "first-12," etc., means.