download the raw data:

http://hgdownload.cse.ucsc.edu/goldenPath/hg16/database/refGene.txt.gz

http://hgdownload.cse.ucsc.edu/goldenPath/hg19/database/refGene.txt.gz

Instead of pissing off UCSC by throwing hundreds of thousands of queries their way, why not just download the various annotation tables (via the table browser or ftp site) and simply process those? That would be vastly simpler than scraping a bunch of web pages.

Don't escape it. You can query their mysql server directly:

mysql --user=genome --host=genome-mysql.cse.ucsc.edu -NB -e 'select * from refGene'

As a former Ensembl team member, I just want to emphasise that scraping websites is absolutely NOT DONE!!! I know of people who were scraping the Ensembl Genome Browser website and were given IP bans because of this (which were lifted again after Ensembl spoke with them and told them how to get the desired data without slowing down / bringing down their production webservers). So, please be aware of this! As already indicated in the other responses, there are many other ways to get the UCSC data (mysql, downloads, Table browser).