Antibiotic resistance genes are often found in structures bounded by copies of IS26, IS257/IS431 or IS1216 that resemble compound (or composite) transposons. However, because of the mechanisms used by IS26 family members, namely that they form cointegrates but cannot resolve them, none of these structures can move together as a coherent single unit. Apparent transposition of these structures is possible via a 2-step process but only if the IS are in direct orientation. An intermolecular reaction catalysed by the IS-encoded transposase and an intramolecular homologous recombination step can occur in either order. In one route, one of the IS bounding the structure forms a cointegrate between the DNA molecule carrying it and a target molecule. Cointegrates formed by either copy-in or targeted conservative routes contain three directly-oriented IS copies and can be resolved by homologous recombination between specific pairs of IS, with one pair leading to apparent transposition of the whole structure. In the other route, homologous recombination first forms a circular intermediate, a translocatable unit or TU, which is incorporated by the transposase either at a random site or adjacent to another IS copy in a target molecule. We therefore conclude that the transposon-like structures are not compound (or composite) transposons and the nomenclature for them should be revised. We propose that the term "pseudo compound transposon" (PCT), first coined in 1989, should be used to describe those structures where the IS are in direct orientation. Structures with the IS in opposite orientation should not be named as transposons.