The zero-field magnetic susceptibilities of Ni(NO3)22H2O, Ni(NO3)24H2O, and Ni(NO3)26H2O have been measured in the liquid-hydrogen and liquid-helium ranges. The dihydrate is obtained by evaporation of a solution at 105°C.1 Its powder susceptibility has a large, sharp, peak at 4.20°K, where it reaches 0.74 cgs/mole, then drops down to 0.2 cgs/mole below 2°K. When measured along the a axis, the susceptibility of monoclinic single crystals of the dihydrate shows a similar peak. It reaches 1.5 cgs/mole, but drops to vanishing values at lower temperatures. The susceptibility in the bc plane reaches only 0.3 cgs/mole, and is nearly isotropic. It drops little below 4.20°K. This behavior is similar2-5 to that of FeCl2 or FeCO3, and suggests the existence of two magnetic sublattices, with strong ferromagnetic interactions within each sublattice, and weaker antiferromagnetic interactions between one sublattice and the other (metamagnetism). Direct evidence for such a layer structure is also provided by a recent x-ray structural work.6 A spin Hamiltonian with S=1 and uniaxial one-ion anisotropy gives results in fair agreement with the experimental data if the exchange interactions are described in the molecular field approximation. The best fit corresponds to g=2.25, D/k=−6.50°K, n1=+0.32 mole/cgs, n2=−2.12 mole/cgs, where n1 and n2 are, respectively, the antiferromagnetic and the ferromagnetic molecular field constants. In the case of the tetrahydrate1 and of the hexahydrate, the powder susceptibility approaches a constant value of 0.35 cgs/mole below 2°K; the data can be fitted to the spin Hamiltonian7 for a nickel ion in a rhombic field, without exchange, with E/k=−2.66°K, D/k=−8.67°K, and g=2.25. A detailed account of this work has been published elsewhere.8