430 Brian K. Hall
2 SIMILARITYHomologous features need not be identical but must share sufficient ‘similarity’
to be recognizable as homologous. Analogous and homoplastic features also are
identified on the basis of similarity of structure or function.^4
Similarity does not necessitate structural equivalence; some features known to
be homologous could not be identified as such on the basis of structural similarity
but can be so identified because we know the phylogenetic history of the feature and
of the taxa that contain the feature, a history that links the feature to homologous
features in earlier taxa. The term and concept,latent homology, has been used for
such features (Box 1).
Homologyis similarity that reflects common descent and ancestry.Homoplasy
is similarity (some might say superficial similarity) arrived at via independent evo-
lution. It is often (usually?) assumed that homologous features share a common
developmental basis. However, different environments or selective pressures can
trigger the appearance of what appear to be similar features in organisms that do
notshare a most recent common ancestor — homoplasy as classically defined —
and where the features therefore develop using different not similar developmental
processes. Parallelism, traditionally one of the classes of homoplasy,canproduce
similar features usingsimilar developmental processes. Furthermore, and there
are many examples, similar (homologous) aspects of the phenotype can arise using
different (nonhomologous) developmental pathways. Homology may or may not
imply conserved development. Homoplasy implies divergent development. Con-
sequently, divergence in developmental pathways is not an adequate criterion to
establish features as homoplastic.^5
Importantly, some features of all organisms share some degree of relationship
and similarity to one another. Shared relationship ranges from: similarity or
even identity of structure in related taxa (exemplified by the limbs of tetrapods),
through similarity of a feature in organisms with a shared ancestor (exemplified
by fish fins and tetrapod limbs), to similarity of a feature in organisms whose last
common ancestor lies deep in metazoan history (exemplified by the tissue cartilage
in invertebrates and vertebrates) and finally to similarity and a shared ancestor
(^4) Determining similarity is not trivial. Gans [1985] provides a valuable evaluation of the in-
terrelationships and potential independence of similarity of structure, development and function.
Also see Haas and Simpson [1946], Riedl [1978], Bowler [1984; 1996], Stevens [1984], Sanderson
and Hufford [1996], D. B. Wake [1991; 1994; 1996; 1999], Hall [1994a; 1999a; 2003a,b], Ospovat
[1995], Bock and Cardew [1999], Butler and Saidel [2000] and Rieppel and Kearney [2002] for
what is meant by similarity.
(^5) See Spemann [1915], de Beer [1971], Riedl [1978], Hall [1994a,b; 1995], Bock and Cardew
[1999] and Svensson [2004] for discussions and examples of divergent (non-homologous) devel-
opmental bases of homologues, Reilly and Lauder [1988] and Hall [2002a; 2003a; 2007a] for
parallelism, and Hall [2003b; 2004] and Hallet al.[2004] for the role of the environment. Hall
[1992] proposed the terms equivalent and non-equivalent developmental processes to reflect con-
served or divergent development. I still prefer these terms to homologous and non-homologous
[Striedter and Northcutt, 1991], if only to avoid confusion over levels of homology. Butler and
Saidel [2000] proposedsyngenyandallogenyfor similar or different genesis.