Spinal muscular atrophy (SMA) is caused by mutations in the SMN1 gene. We have studied the molecular pathology of SMA in 745 unrelated Spanish patients using PCR-RFLP, SMN gene dosage analysis, linkage studies, long-range PCR and direct sequencing. Our systematic approach allowed us to complete genetic testing and risk assessment in 736 SMA patients (98.8%). Females were more frequently affected by the acute form of the disease (type I), whereas chronic forms (type II-III) predominated in males (p<0.008). Absence of the SMN1 gene was detected in 671 patients (90%), and hybrid SMN1-SMN2 genes were observed in 37 cases (5%). Furthermore, we detected 13 small mutations in 28 patients (3.8%), four of which were previously identified in other populations (c.91dupT; c.770_780dup11; p.Tyr272Cys and p.Thr274Ile), while five mutations were found to date only in Spanish patients (c.399_402delAGAG, p.Ile116Phe, p.Gln136Glu, c.740dupC and c.834+2T>G). The c.399_402delAGAG mutation accounted for 1.9% of all Spanish SMA patients. Finally, we discovered four novel mutations: c.312dupA, c.411delT, p.Trp190X and p.Met263Thr. Our results confirm that most SMA cases are due to large genetic rearrangements in the repetitive region of the SMA locus, resulting in absence-dysfunction of the SMN1 gene. By contrast, ancestrally inherited small mutations are responsible for only a small number of cases. Four prevalent changes in exons 3 and 6 (c.399_402delAGAG; c.770_780dup11; p.Tyr272Cys; p.Thr274Ile) accounted for almost 70% of our patients with these subtle mutations. An SMN-SMN dimer model featuring tight hydrophobic-aromatic interactions is proposed to explain the impact of mutations at the C-terminal end of the protein.