The strategic role of SOC in reducing atmospheric CO2 concentrations was recognised in Article 3.4 of the Kyoto Protocol. There are several measures (Recommended Management Practices - RMPs) for sequestering C in the soil, such as reduced or zero tillage, soil application of biosolids (manure, crop residues, compost), cover and deep-rooting crops, etc. Knowledge of the relationship between SOC quality and quantity and soil physical parameters is extremely important to understand the effect of RMPs on soil quality. Total porosity and pore size distribution are among the most important indicators for defining soil physical quality. The effects of 40 years of organic, mixed and mineral fertilisations on pore size distribution were evaluated in a continuous maize system. A long-term experiment has been underway since 1962 on the University of Padova Experimental Farm (Northern Italy). We compared 8 different fertilisation treatments: only organic (L2, farmyard manure - 60 t ha-1y-1; Lq2, liquid manure, 120 t ha-1y-1), only mineral (M2, high mineral input- 300 kg ha-1y-1 N) or mixed inputs (L1M1, farmyard manure - 30 t ha-1y-1+ mineral, 150 kg ha-1y-1 N; Lq1M1 liquid manure - 60 t ha-1y-1+ mineral - 150 kg ha-1y-1 N). Half of the treatments included crop residue incorporation (+ r). The experimental layout was randomised block with three replicates, on plots of 7.8 x 6 m. The soil is a fluvi-calcaric cambisol (CMcf), silty or sandy loam, with sub-basic pH. We investigated pore size distribution by mercury porosimetry. In each plot, two undisturbed samples were collected from the top layer (0-30 cm). Aggregates with a volume of approximately 8 cm3 were air-dried prior to analysis. Pores within the range 10 ìm-600 ìm were analysed with Thermo Finnigan Pascal 140 using wide and ultra dilatometers; pores within the range 0.007 ìm-10 ìm were analysed with Thermo Finnigan Pascal 240 using wide dilatometer. Pore size distribution, classified according to the six classes proposed by Brewer, was analysed with ANOVA; PCA was also applied considering other physical-chemical parameters: texture, pH, stability index, bulk density, non-humic carbon (NCU), 3 humic fractions separated on the basis of the apparent molecular weight: F1 ( >100 kDa); F2 (10-100 kDa); F3 (<10 kDa). Wide variability was observed in the porosity, especially for classes with a diameter of above 30 mm. Treatments affected (p<0.05) only micropores (5-30 mm) and cryptopores (<0.1 mm). M2+r had the highest content of micropores (12.6 % SE ± 1.1 %) and the lowest of cryptopores (1.9% SE ± 0.1). The highest volume of cryptopores was observed in L2 and Lq1M1+R (2.6% SE ± 0.1). Significant correlations were found between pore diameters and SOC: mesopores (30-75 mm) were positively correlated with NCU and F2, cryptopores with F3, the humic fraction with the lowest degree of polycondensation. PCA allowed the most important factors explaining the variability of the system to be identified. The first three components extracted by PCA explained 60% of the total variance. The first component (“organic”; 28% of variance explained) was mainly correlated with organic matter, in particular NCU, F1 and F2, the second (“micro-porosity”; 18% of variance explained) with clay fractions and smallest pore classes (micropores, ultramicropores (0.1-5 mm) and cryptopores) and the third ( “macro-porosity”; 14% of variance explained) with the largest pore classes: mesopores (30-75 mm), macropores (75-100 mm) and ultramacropores (100-600 mm). Treatment effects on the 3 components clearly showed that: 1) the first component separated the mineral from organic fertilisations; 2) L2, M2 and M2+R tended to cluster according to the second component “micro-porosity”; 3) no treatment was clustered by the third component “macro-porosity”.