Previous indications that cloned B virions might be genetically predisposed to generate a particular defective T particle are shown to be inaccurate. T particle generation was found to be a much more random process than was previously believed. We show that the previously observed generation of particular sizes of T particles by B virion pools is due to the random generation of T particles during preparation of first-passage pools of cloned B virions, and these breed true during the additional passages needed to produce visible quantities of T particles. It is also shown that different host cell lines selectively amplify different T particles, suggesting a strong role of host cell factors in T particle replication. Surprisingly, our line of HeLa cells did not generate or replicate detectable T particles of vesicular stomatitis virus (VSV) Indiana after either serial undiluted passage or direct addition of T particles, even though the added T particles strongly interfered with B virion replication. In contrast to VSV, rabies virus generates large amounts of T particles during the first passage of cloned B virions, and every rabies-infected baby hamster kidney-21 cell culture evolves into a persistent carrier state. We find that T particle RNA is biologically inactive although T particle nucleocapsid ribonucleoprotein replicates and interferes in cells coinfected with B virions. Efforts to study the mechanism of T particle generation by in vitro attempts to generate T particles or modify their size (using sheared ribonucleoprotein or chemical or UV mutagenesis) were unsuccessful. The kinetics of UV and nitrous acid inactivation of T particles indicate a smaller target size relative to B virions, even after correcting for lengths of RNA molecules. The intercalating dye proflavine does not photosensitize VSV B virions or T particles when present during replication, indicating that there is little or no RNA base pairing in the helical nucleocapsids of either.