The last few years has seen enormous progress in understanding of protein targeting and translocation across biological membranes. Many of the key molecules involved have been identified, isolated, and the corresponding genes cloned, opening up the way for detailed analysis of the structure and function of these molecular machines. It has become clear that the protein translocation machinery of the endoplasmic reticulum is very closely related to that of bacteria, and probably represents an ancient solution to the problem of how to get a protein across a membrane. One of the thylakoid translocation systems looks as if it will also be very similar, and probably represents a pathway inherited from the ancestral endosymbiont. It is interesting that, so far, there is a perfect correlation between thylakoid proteins which are present in photosynthetic prokaryotes and those which use the sec pathway in chloroplasts; conversely, OE16 and 23 which use the delta pH pathway are not found in cyanobacteria. To date, no Sec-related proteins have been found in mitochondria, although these organelles also arose as a result of endosymbiotic events. However, virtually nothing is known about the insertion of mitochondrially encoded proteins into the inner membrane. Is the inner membrane machinery which translocates cytoplasmically synthesized proteins capable of operating in reverse to export proteins from the matrix, or is there a separate system? Alternatively, do membrane proteins encoded by mitochondrial DNA insert independently of accessory proteins? Unlike nuclear-encoded proteins, proteins encoded by mtDNA are not faced with a choice of membrane and, in principle, could simply partition into the inner membrane. The ancestors of mitochondria almost certainly had a Sec system; has this been lost along with many of the proteins once encoded in the endosymbiont genome, or is there still such a system waiting to be discovered? The answer to this question may also shed light on the controversy concerning the sorting of the inter-membrane space proteins cytochrome c1 and cytochrome b2, as the conservative-sorting hypothesis would predict re-export of matrix intermediates via an ancestral (possibly Sec-type) pathway. Whereas the ER and bacterial systems clearly share homologous proteins, the protein import machineries of mitochondria and chloroplasts appear to be analogous rather than homologous. In both cases, import occurs through contact sites and there are separate translocation complexes in each membrane, however, with the exception of some of the chaperone molecules, the individual protein components do not appear to be related. Their similarities may be a case of convergent rather than divergent evolution, and may reflect what appear to be common requirements for translocation, namely unfolding, a receptor, a pore complex and refolding. There are also important differences. Translocation across the mitochondrial inner membrane is absolutely dependent upon delta psi, but no GTP requirement has been identified. In chloroplasts the reverse is the case. The roles of delta psi and GTP, respectively, remain uncertain, but it is tempting to speculate that they may play a role in regulating the import process, perhaps by controlling the assembly of a functional translocation complex. In the case of peroxisomes, much still remains to be learned. Many genes involved in peroxisome biogenesis have been identified but, in most cases, the biochemical function remains to be elucidated. In this respect, understanding of peroxisome biogenesis is at a similar stage to that of the ER 10 years ago. The coming together of genetic and biochemical approaches, as with the other organelles, should provide many of the answers.