The terms "phenetic" and "phylogenetic" are now common coin in the literature of systematics; but those who use them must beware of accusations that they are doing so in a vague, or private, or otherwise improper fashion. Unfortunately, in the heat of controversy both words have acquired a persuasive usage, pejorative or approbatory according to point of view, and often referring to attitudes rather than actions. This tends to obscure certain very important distinctions which they can draw, and which I propose to examine below. The controversy in this area of taxonomy has involved not one, but several main issues, often confused. (A) How can we reconstruct the historical course of evolution in a credible fashion; which methods are logically and scientifically defensible, and which supply the most credible reconstructions? (B) If we have such information, should it form the basis of our classifications? (C) If it should, what should be the formal structure of such a classification, and what kinds of evolutionary information can it, and should it display? Clearly, (B) and (C) are largely concerned with valuejudgements. They are none the less important for that, but I am not here (and never have been) concerned to argue those points. In my view, the "scientific" content of the controversy mainly concerns (A); and if "phenetic" and "phylogenetic" are to be used in that context, they must therefore denote clearly explicated classes of methods. "Phenetic" has from the outset been explicated primarily in terms of operations that result in a classification; its usage, in the context of (A), to describe the operations and, by extension, the classification is fairly clear (Sokal and Sneath, 1963; Colless, 1967). "Phylogenetic" is more difficult. In its primary sense, it describes a classification, one that reflects or illustrates the course of evolution; but by extension it has come to describe certain operations said to result in such a classification. However, in that original sense, it entails nothing as to precisely how the "genesis" of "phyla" is to be reconstructed. Phenetic methods are not necessarily excluded, nor is a phenetic classification necessarily non-phylogenetic. If we are to discuss methods, in the context of (A) above, we obviously need a less question-begging term for methods that are claimed, rightly or wrongly, to be nonphenetic and to result in a phylogenetic classification. My own solution is to use "phylogenetic," i.e. in quotes, for such methods and, by extension, the classifications that ensue. Any other solution seems bound to draw heavy fire, particularly from those who find comfort in clouded issues. Using some such terminology, we can then discuss in a more meaningful fashion whether, for instance, a given "phylogenetic" method is really non-phenetic; and which of two classifications, one "phylogenetic" and one phenetic, is the better phylogenetic classification. At least we might avoid the fallacy of arguing that the "phylogenetic" one is by definition the better. There might also be some use for "phenetic," to refer to classifications said to result from phenetic operations, but defined by other criteria; though I suspect that its use would be mainly in indicating misuse. In my own usage, which I draw from methods that have been both advocated on theoretical grounds and illustrated in practise; "phenetic" refers to a classificatory process that explicitly or implicitly makes use of the following crucial notions: unit characters; equal (or "non-phylogenetic"see below) weighting of characters; and overall resemblance; all of which-have been discussed elsewhere (Colless, 1967). By "phylogenetic," then, I would refer to a process whose operations actually or ostensibly imply rejection of any of the above