The mechanism of 4p 1P1→4p 3PJ intersystem crossing (ISC) following excitation of the 4p 1P1 level of matrix-isolated atomic zinc is investigated using a pair potentials approach. This is achieved by extending earlier ISC calculations on the Zn⋅RG2 and Zn⋅RG3 complexes to the square planar Zn⋅RG4 and square pyramidal Zn⋅RG5 species which are the building blocks of the Zn⋅RG18 cluster used to represent the isolation of atomic zinc in the substitutional site of a solid rare-gas host. ISC predictions in these clusters are based on whether crossing of the strongly bound 1A1 states, having a 4p 1P1 atomic asymptote, occurs with the repulsive 3E states correlating with the 4p 3PJ atomic level of atomic zinc. Predictions based on 1A1/3E curve crossings for 3E states generated with the calculated ab initio points for the Zn⋅RG 3Σ(pz) states do not agree with matrix observations. Based on similar overestimation of ISC in the Zn⋅RG diatomics, less repulsive Zn⋅RG 3Σ(pz) potential curves are used resulting in excellent agreement between theory and observations in the Zn–RG matrix systems. 1A1/3E curve crossings do not occur in the Zn–Ar system which shows only singlet emission. Curve crossings are found for the Zn–Xe system which exhibits only triplet emission. The Zn–Kr system does not show a crossing of the body mode Q2, which exhibits a strong singlet emission at 258 nm while the waist mode Q3, does have a crossing, resulting in a weak singlet emission at 239 nm and a stronger triplet emission at 312 nm. The efficiency of ISC is determined from Landau–Zener estimates of the surface hopping probabilities between the 1A1 and the 3E states. Differences in the application of this theory in the gas and solid phase are highlighted, indicating that the rapid dissipation of the excited-state energy which occurs in the solid must be included to obtain agreement with observations.