A novel in vivo-model composed of small juvenile rainbow trout and a fish-pathogenic virus is suggested to analyze delivery and antiviral effect of formulated siRNAs. This model was used for testing delivery of intraperitoneally injected siRNAs formulated in polycationic liposomes. These, and to a lesser degree naked siRNAs, primarily entered free intraperitoneal cells including macrophage-like cells. Furthermore uptake correlated with antiviral activity seen as reduced mortality of fish challenged with VHSV. Protection at the disease level was not dependent upon which one of three tested siRNAs was used and protection correlated with up-regulation of an interferon-related gene in the liver indicating a systemic interferon response. The results show the validity of the fish model for testing delivery and non-specific effects of siRNAs in a high throughput vertebrate model. The purchase of chemically synthesized siRNAs is expensive why the use of in vitro transcribed siRNAs was initially tested in fish cell culture. Transfection with three different in vitro transcribed siRNAs specific to the viral glycoprotein gene of the target-virus efficiently inhibited viral multiplication in infected cell cultures, while two of three corresponding control siRNAs, containing four mismatches compared to the target, did not have this effect. This suggested specific interference, but similar results were obtained when the same siRNAs were tested against a heterologous virus. Further analyses revealed that the siRNAs induced a non-target-specific anti-viral effect, which correlated with an upregulation of the interferon induced Mx gene. Accordingly inclusion of a heterologous virus as target control was essential for verification of the specificity of siRNA-induced interference with virus multiplication. Current work on studying the action of chemically synthesized siRNAs designed to suppress expression of the surface glycoprotein G and the large polymerase L of the rhabdoviral target virus is also presented. The results emphasize the use of controls, chice of target gene, type of siRNA and the compromise in using transfection reagents for improved uptake of siRNAs, where these reagents also increase the risk of the siRNAs ending up in a cellular compartment in which stimulation of non-specific anti-viral defence mechanisms will be initiated.