A series of volumes devoted to molecular immunology will contain, for the most part, articles which attempt to explain immunological phenomena in terms of the behavior and properties of particular molecules. Many of the articles in this volume do this. At the same time, there are many instances-and this is particularly so in the case of immunology-where phenomena must first be described and interpreted in terms of the properties and behavior of cells. Most of us would hope that in due course a fuller understanding will be forthcoming. This volume starts off with such a contribution. Perhaps the most fascinat­ ing problem in immunology is how diversity is generated. There are two broad proposals: (1) that complete information exists ab initio (the germ-line theory), and (2) that there is initially a limited amount of information, and diversity is generated by somatic mutation. The issue is unresolved, but Cunningham has taken many of the data which have previously been used to support the germ-line theory and shows that the interpretations are not always clear-cut and can frequently be used to support another possibility-that new specificities may arise after stimulation of appropriate cells by antigens. And he has produced experimental evidence to support this notion. On the other hand, there can be little doubt that to a considerable degree the specificity of the immune response is determined by the selection by antigen of cells with receptors of appropriate specificity. This is essentially a surface phenomenon.