As every chemist knows, there is a direct (if complex) relationship between the molecular structure of a compound and its chemical behavior. Predicting such behavior is possible by an abstract representation of its structure in terms of chemical similarity parameters, socalled 'descriptors'. These are most useful in predicting the pharmacological properties of drug candidates, but are also used in predicting reactivity, toxicity and other important chemical characteristics.

The number-one reference on the topic now contains a wealth of new data: The entire relevant literature over the past eight years has been pain-stakingly surveyed, resulting in hundreds of new descriptors being added to the list, and some 3,000 new references in the bibliography section.

Volume 1 contains an alphabetical listing of around 3300 terms for the chemoinformatic analysis of chemical compound properties, while the second volume contains 6343 references selected from 450 journals with about 7000 authors quoted covering the period from the beginning of molecular descriptor research until the year 2008.

In this second edition, several sections have been completely rewritten and organized in a more effective way. The greatly expanded introductory section now contains several "walk-through" reading lists of selected keywords to make the data even more accessible for novice users.