beta-Lactams are the largest antibiotic family, but are readily compromised by resistance. The result has been a cat-and-mouse game between chemists and bacteria, with the compounds repeatedly modified to overcome emergent resistance. With penicillins, it is possible to obtain spectrum, or beta-lactamase stability, but difficult to combine both. In general, it is better to protect a labile penicillin with an inhibitor, though this strategy is limited by the absence of good inhibitors of AmpC beta-lactamases. Combining spectrum and beta-lactamase stability proved easier with cephalosporins, but it is difficult to cover enterobacteria, anaerobes, non-fermenters and staphylococci with a single compound, and enterococci are consistently resistant. Carbapenems allow the broadest spectrum of available beta-lactams. Less equal or predictable than initial spectrum is how rapidly resistance emerges. This point is especially important pertinent to beta-lactamases; PBP changes compromise all beta-lactams. Spread of plasmidic beta-lactamases destroyed the utility of penicillin G against staphylococci and that of anti-gram-negative penicillins against enterobacteria. Resistance to 'beta-lactamase-stable' cephalosporins has recently spread in enterobacteria, mediated by hyperproduction of AmpC beta-lactamases and extended-spectrum TEM and SHV types. Carbapenems were launched shortly after 3rd-generation cephalosporins, but beta-lactamase-mediated resistance has emerged more slowly. Nevertheless, recent reports of zinc carbapenems in gram-negative bacteria from Japan are disturbing.