Two $\beta$-glucosidase genes of $\mbox{\underline{Cellulomonas}}$ $\mbox{\underline{fimi}}$ ATCC484 were cloned in $\mbox{\underline{Escherichia}}$ $\mbox{\underline{coli}}$ JM83 using plasmid pUC9 as a vector. The gene contained in a constructed plasmid pCF85 encoded enzyme which hydrolyzed both p-nitrophenyl -D-glucopyranoside (PNPG) and cellobiose, while the gene contained in pCF115 produced enzyme which hydrolyzed only PNPG. Southern hybridization indicates that the genes contained in the two plasmids, pCF85 and pCF115, are not identical. Since the enzyme activity originated by the two genes depended on the orientation of the chromosomal inserts in the vector plasmid, the genes seem to be under the control of the lacZ promoter of pUC9 in $\mbox{\underline{E.}}$ $\mbox{\underline{coli}}$. Since I have been interested in an enzyme which hydrolyze cellobiose, my further study was concentrated on the gene contained in pCF85. The size of the $\mbox{\underline{C.}}$ $\mbox{\underline{fimi}}$ chromosomal insert in the recombinant plasmid pCF85 was found to have a strong effect on the production of $\beta$-glucosidase. Through a sequential subclonings, the size of the chromosomal insert was reduced to 1.8 kb and constructed a new plasmid pCF18. The enzyme encoded by the gene in pCF18 cleaved cellobiose into two molecules of glucose, and thus, the gene cloned in pCF18 was identified as a true $\beta$-glucosidase gene. The enzyme from crude extract of $\mbox{\underline{E.}}$ $\mbox{\underline{coli}}$ (pCF18) was purified. SDS-polyacrylamide gel electrophoresis of these final preparations revealed that molecular weight of this enzyme was 56,000 dalton. And analytical gel filtration suggested that this enzyme consists of a single polypeptide. Inhibition caused by heavy metals and activation by dithiothreitol suggest the existence of essential thiol group in the enzyme. The fact that an enzyme solution containing dithiothreitol showed increase in heat stability also su...