The classification of the various birds that make up the order has long been controversial and difficult, particularly in the case of the nightjars. Most taxonomists consider the following families, but some may be polyphenetic in their own distinct orders:
Traditionally, they were regarded, on morphological grounds, as being midway between the owls (Strigiformes) and the swifts. Like the owls, they are nocturnal hunters with a highly developed sense of sight, and like the swifts they are excellent flyers with small, weak legs. At one time or another, they have been allied with owls, swifts, kingfishers, hoopoes, mousebirds, hornbills, rollers, bee-eaters, woodpeckers, trogons and hummingbirds.
Based on analysis of DNA sequence data – notably β-fibrinogen intron 7 – Fain and Houde considered the families of the Caprimulgiformes to be members of the proposed clade Metaves, which also includes the hoatzin, tropicbirds, sandgrouse, pigeons, kagu, sunbittern, mesites, flamingos, grebes and swifts and hummingbirds. This clade was also found by the expanded study of Ericson et al. (2006), but support was extremely weak.
While only the latter study recovered monophyly of the Cypselomorphae (see below) within Metaves, the former was based on only a single locus and could not resolve their relationships according to standard criteria of statistical confidence. No morphological synapomorphies have been found that uniquely unite Metaves (or Caprimulgiformes for that matter), but numerous unlinked nuclear genes independently support their monophyly either in majority or whole. Ericson et al. (2006) concluded that if valid, the "Metaves" must originate quite some time before the Paleogene, and they reconciled this with the fossil record.
While the relationships of cypselomorphs are a subject of ongoing debate, the phylogeny of the individual lineages is better resolved. Much of the remaining uncertainty regards minor details.