Abstract:
1-1 is torically, the predominan t method of reconstructing
phylogenies has been through the use of morphological characters.
There are new techniques now gaining acceptance, including molecular
techniques al1d chromosomal information.
Altl10ugh the study of behaviour has been used in a comparative
framework, these analyses have, historically, been based on intuition.
Hennig (1966) devised a neV\' method of reconstructing phylogenies
which provided a 110ncircular method for formulating, testing and
refining phylogenies. Subsequent s)Tstematists had virtually abandoned
ecological and beha\lioural data as primary indicators of phylogenetic
relationships (Brooks and McLennan 1991). Therefore, in a modern
cladistic framework (sensu Hennig) the analysis of behavioural traits
remains underrepresented as a method of reconstructing phylogenies.
This thesis will reconstruct the phylogeny for species of black flies
(Diptera: Simuliidae), using two steps. The first step is to thoroughl)'
understand and explain the cocoon spinning in black fly larvae. There
have bee115 previous descriptions of cocoon spinning, but all were
incomplete or erroneous. The advances in technology, including video
recorders and VCRs, have allowed this behaviour to be analyzed in great
detail in 20 different species. A complete description of the cocoon
spinning of Simulium \littatum is given. This description will be used as
a template for the other species observed.
The description and understanding of cococ)n spinning was the first
step in undertaking a phylogenetic analysis using this behaviour. The
behaviour was then broken down and analyzed, revealing 23 characters,
3
either qualitative and quantitative in nature. These characters were
assessed in a cladistic framework (sensu Hennig) and a phylogenetic tree
was reconstructed with a e.I of 0.91 and an R.I. of 0.96. This
phylogenetic tree closely resembles a previously established
pllylogenetic tree produced from morphological and cytological
information.
The importance of this result is the indication that, contrary to
some authors, behavioural characters, if used properly, can add very
informative characters to a data set.
