Two novel gas-tight snow samplers (snow-can and snow-tube) are presented and the performance of the snow-can in a field trial was assessed. The methodology for the sampling, extraction and analysis of persistent organic pollutants (POPs) are detailed. These samplers allow the various components of a snow sample to be analysed separately; these included the meltwater (MW), particulate matter (GFF) and vapour in the headspace (HS). Snow samples collected on the Punta Indren glacier in the Italian Alps revealed the occurrence of polychlorinated biphenyls (PCBs), polycyclic aromatic hydrocarbons (PAHs) and organochlorine pesticides (OC). Replicate samples of the same snow type were undertaken as a test of sampling precision. Relative standard deviations (RSDs) for ∑PCBs and ∑PAHs were 30% and 35% respectively. The lowest precision was found for the particle-laden snow, notably for the heavier PCB homologues. For the chlorinated compounds, the pesticides lindane and endosulfan-I had the highest levels in snow, with mean concentrations of 402 and 103 pg l−1 (snow meltwater) respectively. The vapour present in the headspace (HS) comprised a minor component of a collected sample for all compounds, but HS concentrations for three lighter PAHs gave good agreement with those calculated based on their dimensionless Henry's law constants. This suggests that volatilisation during melting of aged snow-can be reasonably predicted with knowledge of the temperature-dependent Henry's law constant.