In spite of obvious and frequently observed morphological variation in geog
raphic populations of sand flies, taxonomists are reluctant to create subsp
ecies. In doing so they avoid confusion caused by the accumulation of new t
axa, but they also obscure natural processes, some of them related to proba
ble species emergence. We investigated here how far the partitioning of siz
e, shape and form could be a relevant technique for exploring early evoluti
onary processes. We used a typical case of species, subspecies and cryptic
species among three related taxa: Lutzomyia carrerai, its subspecies L. car
rerai thula and L. yucumensis, a cryptic species of L. carrerai. Univariate
and multivariate metric analyses showed global size differences, even betw
een sympatric taxa (L. carrerai and L. yucumensis). When allometric effects
were discounted, the subspecies L. c. thula still disclosed important diff
erences from its putative type species, more pronounced than those existing
between the two cryptic species, L. yucumensis and L. carrerai. A similar
result was obtained after removing isometric size. These analyses question
the validity of L. carrerai thula as a subspecies, and show that morphometr
ics may be a useful tool for separating simple geographic variation from ev
olutionary divergence. (C) 2000 Academie des sciences/Editions scientifique
s et medicales Elsevier SAS.