In examining the problem of conflicting views about mechanical and adaptive functions and phylogenetics, I will not argue that some degree of functional knowledge in the broad sense is important for any causally meaningful understanding of character ordering and polarity to estimate evolutionary history. To me this is obvious. Taxograms based on parsimony analysis, as Mahner and Bunge (1997) called them, are only "pretheoretical classifications", with their studied emphasis on pretheoretical, not phylogenies. But in reading the literature related to functional analysis and its connection to phylogeny, it is also obvious, as alluded to previously, that even students who are deeply convinced that functional explanations must be somehow part of phylogenetic explanations have conflicting views on the very concepts of function: what is and what is not "scientific" in functional biology, and how functional studies bear on phylogenetic explanations.

There are usually obvious levels of resolution, both for adaptational and phylogenetic assessments, and these will vary according to the nature of available evidence. As I see it, there are, minimally, two important and related questions regarding the significance of adaptational analysis and its relationship to phylogenetics: How reliable and independent are functional, adaptational, and phylogenetic estimations, and can they be truly independent?; and how is phylogenetic analysis connected theoretically, and therefore methodologically, to functional biology (sensu lato)?

It is my position that the feedbacks in a rigorously comparative framework between form, function, adaptive appraisal, and phyletics of features (the temporally-looped relationship noted previously) and the full consideration of the fossil record make for robust hypothesis testing. Employing holistic Darwinian explanations for character complexes is far more scientific, in a biological sense, than any Popperian, deep-time-deprived, and structuralist pursuit of morphology by itself can be. With this combinatorial analysis of extant and fossil data and with a full immersion in the most current understanding of developmental biology (including developmental genetics), if the level of analysis demands or permits, homologies can be distinguished from homoplasies with confidence.

We may ask now, in what specific way is functional analysis conducive to phylogenetic estimation? I would first like to reject the theoretical position inherent in often-cited cover statements to the effect that an investigator "will consider function whenever it is useful beyond `character' analysis" or "functional information is merely more grist for the mill" in the construction of ephemeral data bases for parsimony analyses heading for "consensus," as both of these positions are sometimes stated and implied (e.g., Simmons 1993; I will not expand the discussion here on the obvious that a science of evolutionary history that aims to distance itself from a causally meaningful testing of phylogenetic hypotheses is a greatly diminished one).

Quite simply, as originally outlined by Bock (1981), functional-adaptive analysis is character analysis aiming to establish the reliable homologies, be they synapomorphies or transformational homologies with highly corroborated directions and sequences, as distinct from homoplasies. This is fundamentally a Darwinian approach to phylogenetics, because it is a selection-related evaluation of features in an evolutionary context. In fact, the foundations of this view were clearly enunciated by Darwin. Given the geological and geographical contexts and an understanding of developmental homologies, functional-adaptive analysis of fossils and living forms (the latter the designated models that are researched in detail; Szalay and Sargis, in press) can inform with often great confidence about homologous versus homoplasious features in the comparison of details of functioning complexes that perform an adaptive function.

As one example, in a recent study Szalay and Schrenk (1998) compared the living "edentate" and digging mammals (the xenarthran Cingulata and Myrmecophagidae, and the Pholidota), with fossil "edentates" such as palaeanodonts and the Messel Eocene Eurotamandua. The numerous function-driven polarity determinations of this study of the enigmatic phylogenetic relationships involved a high-probability adaptive framework, specifically that the fossils were also myrmecophagous and had obvious digging adaptations, either to get at food or shelter. Going beyond the well-established similarities relating to the myrmecophagous and digging habitus, it was concluded, based on the disparate form-function strategies in achieving that way of life, that the various clustered "synapomorphies" derived from parsimony analysis (most of these being previously poorly assessed adaptive or merely verbal "similarities") of the living species of the American and Old World groups were not homologous by either structural or functional-adaptive criteria. Consequently, they are not indicative of ties between the palaeanodonts and Eurotamandua, on one hand, and pangolins, on the other. The inferred adaptational background for the fossils and their structural-functional assessment in that light were fundamental in reaching these taxic conclusions. Homoplasies could be rejected prior to taxic analysis. Subsequently, corroborated phylogenetic trees fall out of such a character analysis. The Darwinian method is certainly not the straw man characterized by claims of cladists regarding functional-adaptive analysis (e.g., Schuh, 2000) that "only adaptive characters are used in phylogenetics" by Darwinians. In fact, it is the incidental and phyletically informed attributes in addition to the adaptive solutions that are sought to add to a data base as either shared and derived or transformationally revealing attributes.

The alternatives to a Darwinian analysis of morphological attributes, particularly the cranioskeletal characters that vertebrate paleontologists rely on and that I am most familiar with, usually offer a method through which a mixed suite of similarities are subjected to parsimony procedures to sort out homoplasious and homologous similarities. The evolutionary vacuum (i.e., a lack of theoretical justification) within which, I believe, this method is practiced has been discussed elsewhere (Bock 1981; Szalay 1994). But beyond the general weakness is the complete elimination of the explicit research-based phyletics of the characters used. Phyletics, which includes both the ordering and polarization of characters, is conflated with taxic analysis, as often there is nothing beyond taxic sister-group arrangements that might give justifiable indications of a direction of transformations of non-identical (hence, not truly synapomorphous) traits. Stratocladistics also endorses a parsimony-driven cladistic approach but adds a stratigraphical dimension that improves the reliability of its taxograms (see Fisher 1994). It invariably appears that parsimony-based cladistics happens without attempts to independently test postulated synapomorphies.

The whole enterprise of phylogenetic estimation or analysis is a highly probabilistic one. If this is accepted - and I believe that it should be - there can be no theoretical objections to a procedure that is based on the guidelines provided by the mechanical and adaptational analysis of living and fossil species and their specific aspects. Most important, adaptational assessment sets a framework for a character analysis that is also necessarily transformational, as we need to decide rationally (based on sound theory) the ancestral and derived conditions. Such character analysis permits the rejection of homoplasies and allows for ancestor-descendant or branching level-specific corroboration of homologies for taxic analysis, based on some biologically and paleontologically meaningful probability assessment. Taxon phylogenies fall out of such Darwinian analysis and render the former far more probable than an axiomatized and abiological parsimony search for the congruence of often atomized traits of various taxa.