How to perform a chromosomal walk
I believe that the first publication to describe this approach was Bender et al., Science, 1983. In that paper they explain how they "walked" along the chromosome to reach the Bithorax Complex (BX-C).
This approach requires two reagents:
a cloned DNA fragment that you have mapped to a defined genetic interval, and possibly to a cytogenetic interval as well (this is only possible if the organism in question has cytogenetic tools).
a genomic library of larger cloned DNA fragments.
Step 1. Use the DNA fragment to screen the library and isolate genomic clones.
Step 2. Purify each genomic clone and prepare enough DNA to construct a restriction map of the foreign insert (N.B., Bender et al., actually used heteroduplex mapping--an EM-based technique--as well) in each clone.
Step 3. Isolate a suitable DNA fragment, devoid of repetitive DNA, from one end of the clone that extends the farthest along the genome in the direction in which you wish to "walk".
Step 4. Congratulations, you have taken one "step." Now use the new DNA fragment as a probe and screen the genomic library again. Each time you screen the library you will recover all of the genomic clones that hybridize, or overlap, with the fragment you are using as a probe.
Step 5. --> go back to Step 2. above.
Each time you take another "step" you are isolating cloned DNA that is taking you closer to your target (the gene of interest). For the very first step you will not know which clone is extending in the desired direction (typically) so you need some type of complementary assay to decide which end of the overlapping clones' restriction map is the direction you want to pursue. In other words, you need to align the physical map to the genetic map. In Drosophila one can use in situ hybridization to polytene chromosome squashes to visualize where the genomic clones are derived from in the genome, and this lets you pick the correct clone at that first step.
Periodically during the walk you will need to test the new genomic clones by in situ to find out when you are getting close to your target (or if you have even walked past your destination).
Ideally you will have some chromosomal rearrangements that alter the physical and genetic maps, and let you unambiguously determine when the chromosomal walk has reached, or entered, the locus. Alternatively, a strong loss-of-function mutation in your target gene might be a deletion or deficiency that causes a RFLP, and isolating genomic DNA from such a mutant will let you use a Southern Blot to confirm that you have reached the gene.