Brain amyloid plaques are a hallmark of Alzheimer's disease (AD), and primarily consist of aggregated Aβ peptides. While Aβ 1-40 and Aβ 1-42 are the most abundant, a number of other Aβ peptides have also been identified. Studies have indicated differential toxicity for these various Aβ peptides, but in vivo toxicity has not been systematically tested. To address this issue, we generated improved transgenic Drosophila UAS strains expressing 11 pertinent Aβ peptides. UAS transgenic flies were generated by identical chromosomal insertion, hence removing any transgenic position effects, and crossed to a novel and robust Gal4 driver line. Using this improved Gal4/UAS set-up, survival and activity assays revealed that Aβ 1-42 severely shortens lifespan and reduces activity. N-terminal truncated peptides were quite toxic, with 3-42 similar to 1-42, while 11-42 showed a pronounced but less severe phenotype. N-terminal mutations in 3-42 (E3A) or 11-42 (E11A) resulted in reduced toxicity for 11-42, and reduced aggregation for both variants. Strikingly, C-terminal truncation of Aβ (1-41, -40, -39, -38, -37) were non-toxic. In contrast, C-terminal extension to 1-43 resulted in reduced lifespan and activity, but not to the same extent as 1-42. Mutating residue 42 in 1-42 (A42D, A42R and A42W) greatly reduced Aβ accumulation and toxicity. Histological and biochemical analysis revealed strong correlation between in vivo toxicity and brain Aβ aggregate load, as well as amount of insoluble Aβ. This systematic Drosophila in vivo and in vitro analysis reveals crucial N- and C-terminal specificity for Aβ neurotoxicity and aggregation, and underscores the importance of residues 1-10 and E11, as well as a pivotal role of A42.