Fitch (in Keim et al (1981) until I find the better ref) proposed a test: consider:

with present day sequences at the top. Can test by adding more species to the tree (dashed lines below for the newly added data. If we're dealing with divergence, then the predicted ancestors (at the first nodes below) will be closer than the two initial sequences; if however we're dealing with convergence, then the ancestors will be more distant.

Good evidence for convergence is rare; best case I've seen is Stewart & Wilson (1987). They found langurs have 10 amino acid substitutions in lysozyme during the same period when baboons have only 4. More remarkably 5 of the 10 are shared with the cow yet langurs are primates. They argue that this convergence is a response to dietary selection with the langur eating leaves and having similar digestive constraints to ruminants.

When look at many cloned genes, now obvious that within species there are great similarities for different genes. ß globins in human are an example. If we have a cluster of similar genes, then the cluster provides opportunities for further duplication and for loss of genes by misalignment and crossing over. The resultant hybrid genes are more similar to the ancestors than either ancestor is to each other. This process - gene conversion - can occur even when the gene number is unaltered. The big issue is how to detect it.

Need to do piecewise comparisons involving inter- and intra-specific alignments. An example from our lab is that rats have multiple ß globin genes (probably 4) and mice have 2 (called ß-major and ß-minor). We have sequenced much of the genes of the rat and begun the piecewise alignments.

Stolzfuss & Milkman (1988) examined the nature of genetic variation in DNA of populations of E. coli. They find that clonal segments of DNA appear to be replaced not individual bases. This makes good sense as a circular DNA requires two breakpoints for recombination. It also appears that this mechanism may apply to natural populations of many bacteria.

More remarkable are similar observations for DNA exchange across conventional species and genus lines in bacteria. Mazodier & Davis (1988) have reviewed the observations to date. The data suggest that even distantly related bacteria can exchange genetic material although the phenomenon is normally detected only under strong selective pressure. The rate may be limited by the lack of sequence similarity to drive efficient genetic recombination. In this sense all bacteria can be thought of as one panmictic species!

The relation between phylogenetics and cladistitics