Show the source code or download the program by clicking on this link: tax7.for
This is a FORTRAN II program that I wrote in 1966 to infer a tree using Camin-Sokal parsimony with characters that had "character state trees". It read in a data set that included a set of character state trees, where there could be up to 9 states (0 to 8) with state 9 being the "unknown" state. I sent the original program to Bob Sokal and Jim Rohlf and they reported that it ran faster than their program. That is not surprising since their program wrote out information to disk or tape and then read it back in.
This program is not the original version but is the program after a little development. It uses standard FORTRAN II but also uses a nonstandard operation (&) which is a bitwise AND of the bits in two integers. That was available in the FORTRAN II compiler for the IBM 7094 at the University of Chicago.
The program reads in data that has up to 70 characters, each with up to 9 states (0 through 8). State 9 is used to indicate
that the state us unknown. Each species has a name of 10 characters long. After the initial line, which has the number of species
and the number of characters, there is one line for each character which indicates the "character state tree" for that character.
In each of those lines are 2 characters indicating which state is the immediate ancestor of that state on the character state
tree. They are respectfully for states 0, 1, 2, ..., 8. The 9th pair of characters is 0 if the character is to be ignored
in this run, otherwise something else. If a state is the ancestral state the number for that state is 10. Thus for a character
state tree that has ancestral state 1, with state 0 arising from that, state 2 also arising from state 1, and state 3 arising
from state 2, the line for the character will be
110 1 2-1-1-1-1-1 1
(the -1 entries showing that the state is not used for that character).
The program can handle up to 40 species, but that number is easily increased.
The algorithm for tree construction is sequential addition of species. After the last species is added, there are then multiple rounds of Nearest Neighbor Interchanges which end when one of them does not change the tree.