Partial warps, phylogeny, and ontogeny: a comment on Fink and Zelditch (1995).

Fin k and Zeld itch (1995) recen tly prop osed a m ethod by w hich one m ay infer phylo geny from onto genetic tran sform a-tions of continuous m orphological data. They coded ontogenetic reg ression s of sh ap e on size to obtain discre te characters, w hich they used as input into a parsim ony ana lysis. A s sh ape variab les, they used p ar tial warp scores obtained from an an al-ysis based on thin-plate splines (TPS), w hich, they had prev iou sly arg ued, could be considere d hom ologous characters and could be used to determ ine sy na pom or-phies on a phylo genetic tree (Z eld itch et al., 1995). Here we discuss several asp ects of their protocol that call into question their assert ion that the ontogeny of sh ap e (as represen ted by discrete charac teriza-tions of reg ression s of par tial warp scores) is a suitable charac ter set for phy logeny es-tim ation. A sim ilar note (R ohlf, 19 98) discusses more theoretical objections to their m ethodology, and challeng es the notion that par tial warp scores viewed sep arately are biolo gically hom olo gou s and interp ret-able charac ters. Additiona lly, we w ill address two issues raise d by Fink and Zeld-itch (199 5). First, do es the choice of a different star ting form (referen ce) result in a differen t phy lo geny ? Second, does the choice of an altern ative basis for tang ent sp ace resu lt in a differen t phylogeny ? The pro to col of Fink and Z eld itch (199 5) is as follow s. First, a reference con® gu ration is generat ed as the m ea n of several juvenile sp ecim ens from a chosen outgrou p sp ecies. Fink and Zeld itch cho se the average of three juven ile specim en s from one of their outgrou ps, Pygopristis denticulata, a s t h eir r e fer en ce s p e cim en (h er e a ft er c a lle d t he r e fe ren c e ou tg r ou p). T h ey alig ned the specim en s using sha pe coordinates (Bo okstein, 1986) , choosin g the long axis …