A study has been made of the factors and mechanism leading to appearance of the so-called EcoRI activity described by Polisky et al. (1975) in the restrictase EcoRI preparations. The preparations of purified restrictase EcoRI, precipitated at 0.9 ammonium sulphate saturation, as well as that obtained using standard techniques have been found to contain an admixture of an endonuclease which at neutral pH and high ionic strength multiply cleaves those DNAs which normally have only one recognition site for EcoRI. Under the standard conditions for EcoRI digestion this activity is found only when large amounts of freshly isolated enzyme are added to the incubation mixture and it is sharply enhanced by replacement of Mg2+ with Mn2+. The number and size of DNA fragments produced under such conditions practically do not differ from those found under the so-called EcoRI conditions, that is for alkaline pH values and low ionic strength. The optimum incubation mixture for the EcoRI activity has been found to be 10 mM Tris . HCl buffer (pH 8.8) + 2 mM Mn2+. Similar activity is induced also by addition to EcoRI solution of 40--50% glycerol or a number of organic solvents (dimethylacetamide (DMA), dimethylformamide (DMF), dimethylsulphoxide (DMSO), sulphalane (SP) in concentrations from 1 to 6%. The EcoRI activity induced by 50% glycerol or at alkaline pH values and low ionic strength is suppressed or sharply inhibited by 2--3 mM parachloromercuribenzoate (PCMB), while EcoRI is not sensitive to this agent. The DNA fragments cleaved by EcoRI have cohesive termini and can be easily ligated. It is suggested that the EcoRI activity can be due not only (or largely not) to modification of the "recognizing capacity" of the EcoRI restrictase but not activation of a latent specific endonuclease which is present in the restrictase preparation as an impurity.