There are two types of bootstrap values in common use: Felsenstein confidence and transfer expectation.
Bootstrap values are calculated by comparing a tree against a resampled ensemble of trees.
Boostrap values are typically represented as integers in the range 0...100, or as fractional values 0.0 to 1.0. They may also be real-valued percentages with decimals, or in rare cases you may see an integer which is the number of replicates where the edge was reproduced. If there are exactly 100 replicates in the tree ensemble, this is a percentage, but otherwise it's not a percentage and this can be very confusing if you're not aware of it.
There should be one bootstrap value for every edge. However, trees in Newick format store bootstrap values as node labels rather than edge labels, for the simple reason that Newick does not support edge labels. This causes at least one missing value because there are fewer internal nodes than internal edges. Just one of many problems with the Newick file format, but it is so widely used that we're stuck with it.
