Two kind of data are available: web site pages download data and ftp ones.

     Analyzing the first ones is of interesting for the mirror site to calculate the penetration of its own diffusion strategy, and of the single services offered. But it is not relevant on the NLM side: NLM web have a  lot of  well knowed services, including Medline and some others, and a click on VHP should be also done by uninterested users.

     The analisys of the second kind of data are very important for two reasons: the study of the use month by month of the download service, and a view of what is downloaded.

      In this paper we have done this work for the "VISIBLE HUMAN DATASET - MILANO MIRROR SITE^®".

     First of all we have analyzed VHD-MMS web data collections, from its start in July, 1997. This work has showed, month by month, if Milano Mirror Site was already well knowned, or if, and where, we had to set up other initiatives to divulgate the existence of the dataset mirror site. A complete analisys of the pages retrieved has showed also if newest services were of interest, or not.

     FTP data analisys has been done in two modalities. First we have studied VHD-MMS FTP data collections to put out two kinds of results: from where VHD-MMS FTP site were used, and for downloaded what kind of data.

     Second we have studied NLM FTP site data collections over the same restricted period (5 months), in the same two modalities done for VHD-MMS ones. Then we have put ones in front of others.