An infectious nucleic acid has been extracted from the K (Kilham) rat virus. The virus, grown on cultures of secondary rat embryo fibroblast, was first purified by the following method. After freezing, thawing and low-speed centrifugation, resuspended pellets were homogenized, subjected to ultrasonic vibration, and treated with receptor-destroying enzyme followed by trypsin and deoxycholate. After a second low-speed centrifugation, virus was finally concentrated by high-speed (25,000 rev./min) centrifugation over a potassium tartrate cushion. DNA was extracted by subjecting the purified dialysed virus suspension to osmotic shock with 11 m-urea. After the addition of caesium chloride to stabilize the density at 1.4 g/cm 3, the DNA was purified by density-gradient centrifugation and the resultant fractions assayed for DNA content by ultraviolet absorption. The fractions with the highest o.d. 260: o.d. 280 and o.d. 260: o.d. 230 ratios were pooled and dialysed. Table 1 shows the infectivity of the purified DNA three logarithmic steps lower than the starting purified virus, as measured by plaque formation on rat embryo cells. DNase treatment reduced this infectivity titre 25-fold, but treatment with antiserum chromatographed on DEAE cellulose (to remove native DNase activity) was without effect, although the serum had a plaquereducing titre of 1: 256 when tested against the purified virus. The strandedness of the DNA was investigated by three procedures: acridine orange staining, melting curve and the effect of formaldehyde upon the ultraviolet absorption spectrum. The yellow-green fluorescent colour upon acridine orange staining, the narrow melting range in the region of 87 °C (Fig. 1) and the failure of formaldehyde to alter the ultraviolet absorption spectrum, all indicate that the K virus DNA is double stranded. The buoyant density of the purified DNA obtained by caesium chloride density-gradient centrifugation with phage 2C DNA as marker, was 1.704 g/cm 3 (Fig. 2). The percentage G + C calculated from this value was 45%. This was also in agreement with the value of 43% calculated from the T m value. From the sedimentation coefficient of 12.9 s (Fig. 3), a molecular weight of 2,000,000 was calculated. The purified DNA, and DNA obtained by treatment of purified virus with isoamyl acetate and 8 m-urea, were examined by electron microscopy. In all cases the DNA strands observed were linear, and in no instance was any example of circular DNA found. Figures 4 to 6 represent frequency distributions of strand lengths for the isoamyl acetate, urea and purified DNA preparations respectively. The peak incidence of strand length ranges between 0.6 and 0.8 μ, and the arithmetic mean between 0.75 and 0.86 μ for the three DNA's examined. Because of the risk of producing numerous DNA fragments due to the violent action of isoamyl acetate, only the combined results of examination of the urea and purified DNA materials were pooled; these are presented in Fig. 7, which shows a peak length incidence of 0.83 μ. If one assumes the DNA strands to be in form A, with 2.55 Å separating adjacent bases, then on the basis of a 0.83 μ strand length with 3,250 base pairs, 44% of them being G + C, the molecular weight can be calculated as 2,140,000. If, however the DNA is assumed to be in form B with a distance of 3.4 Å separating adjacent bases, the molecular weight comes to approximately 1,600,000. Both these figures agree with the value of 2,000,000 obtained from the sedimentation coefficient. The DNA of K virus then represents the smallest known molecular weight of a double-stranded DNA genome (Table 2), and with 3,000 base pairs, it should be capable of coding only three to five proteins at the very most.