102 human Bence Jones proteins have been purified by gel filtration, digested with trypsin, and analyzed by peptide mapping. In several cases Bence Jones "fragments", corresponding to the variable half of the corresponding proteins, were observed. The peptide maps of the proteins were compared to establish whether any identical proteins were present in the sample analyzed. No Bence Jones protein showed a peptide map identical to that of any other protein, although remarkable similarities in the peptide maps were observed for some proteins. Two proteins that gave very similar peptide maps were then examined in detail, by purifying and analyzing the tryptic peptides. It was then found that these two proteins differ in amino acid sequence in at least six positions.
The probability of not finding two identical sequences by examining a sample extracted from populations of light chains of different sizes has been calculated. This has led to an estimate of the minimal size of the population of light chain sequences in humans. The number of light chain sequences appears to be at least a few thousand.
Information on the frequency of Inv and Oz antigenic determinants and on the relative frequency of subtypes of K chains has been obtained. Proteins of KI subtype are found most frequently. The possibility that different subtypes may be predominant in different species is discussed in relation to the evolutionary arguments used in favor of the somatic theories on the origin of variability of immunoglobulin chains.