There's actually a very interesting primary source that deals with this. The autobiography You Can't Win by early 20th century reformed thief/crook/hobo/con-man Jack Black. He spent most of his life hopping trains and thieving his way around the western US and Canada between 1880-1910 or so before he went straight and began working for a newspaper in San Francisco. It's not perfectly reliable, since many of the anecdotes are surely exaggerated, and memory fades, but it's a great read and he does provide some detail about swapping names and stories between cities:

Only the large cities attempted anything in the way of identification. The Bertillion system was in the experimental stage and fingerprinting unknown in police work. We jumped from one state to another, kept away from the cities, lived almost entirely on the road except in the dead of winter, and spent our money in the jungles...

This only worked up to a certain point for Jack, for as time went on and his criminal resume grew longer, he had the misfortune of occasionally bumping into people that knew who he was. I cannot say if his experience was typical for a career criminal but I suppose if you live in a particular underworld for long enough, someone's bound to blow your cover to serve an old grudge.

Edit: I just want to take a moment to share the Black's dedication:

...to that dirty, drunken, disreputable crippled beggar, “Sticks” Sullivan, who picked the buckshot out of my back — under the bridge — at Baraboo, Wisconsin.

So you know how good that book is right off the bat.

This was actually quite common and led to the development and widespread adoption of fingerprinting.

When the vast majority of people lived in small communities, and thus the vast majority of crime was local, it was simple to identify and keep track of criminals. With the nascent rise of what would come to be known as "globalization" this changed; a criminal could move from town to town, city to city. And in doing so he could change his name. no longer could a criminal be brought to court and have his longtime friends and companions (or enemies for that mater) identify him with confidence. In one sense this didn't matter--if a criminal was caught, he was sentenced, and it didn't matter what name he went by. But with the advent of new philosophies on criminal justice the concept of recidivism came to the fore. A man who was arrested for a first offense was viewed and sentenced differently than one who was a repeat offender. In order for this variable sentencing structure to be applied it was necessary to know exactly who a criminal was (thereby knowing his criminal history)--not merely to know what crime he had most recently committed.

Many modalities of biometric measurements came to be used to keep track of criminals. One method was to minutely measure multiple parts of a man's body--tip of nose to chin, length of palmar crease, cranial circumference, etc. etc.--and with this catalog of data be able to identify repeat offenders regardless of a name change. This approach was obviously flawed, and eventually fingerprinting was adopted. Only later was fingerprinting used to identify those who had committed a crime--originally it existed solely to positively identify repeat offenders. The development of fingerprinting science, and especially the cataloging system appropriated from Chinese dictionaries, is a fascinating tale in its own right.

Of note: it was not merely criminals who would change their names to avoid identification. Soldiers in the Civil War would often enlist, claim the generous enlistment bounty, and then go AWOL.....only to reenlist in another unit under a different name to claim a second (or third or fourth) enlistment bounty. When caught this was a capital offense.

Theres was the system of Bertillonage, developed in the late 1880s. It was something like an identification via fingerprint or DNA, but obviously not as exact and it couldn't be used to match an individual to a crime scene like the modern techniques can. Quote:

Using Bertillonage, as the method came to be called, a trained expert would take eleven specified measurements, including the length and width of the head, height, and the length of the left foot and several other physical dimensions, and enter them all on a card, or fiche anthropomorphique.7 Police would keep this card with thousands of others, nested in overlapping categories of small, medium, and large for each head length, height, foot size, et cetera. The system was organized so logically and simply that a trained clerk could work his way through the categories and produce a positive ID in a matter of minutes. During the first year of experimental use, Bertillon identified three hundred repeat offenders. By the end of that year, the French prison system introduced the method to all the nation’s penitentiaries, and in 1888 it became mandatory in all the nation’s police stations. Soon police forces throughout the world adopted the system, including those in other Western European countries, India, and Russia. In 1897, it became the official technique of the FBI’s forerunner, the National Bureau of Identification. Later, a method was devised to transmit Bertillon numbers as a code over telegraph. For the first time, a fleeing criminal’s identity would reach police in a foreign country—even across an ocean—before he or she could arrive. “The prisoner who passes through [Bertillon’s] hands is … forever ‘spotted,’ ” wrote the American journalist Ida Tarbell, who visited Bertillon in his lab.8 "He may efface his tattooing, compress his chest, dye his hair, extract his teeth, scar his body, dissimulate his height. It is useless. The record against him is unfailing. He cannot pass the Bertillon archives without recognition; and, if he is at large, the relentless record may be made to follow him into every corner of the globe where there is a printing press, and every man who reads may become a detective, furnished with information which will establish his identity. He is never again safe."

Source: Douglas Starr: The Killer of Little Shepherds. A True Crime Story and the Birth of Forensic Science. The book is about Joseph Vacher, a vagabond who murdered and raped people in southern France in the late 1890s. He managed to evade capture for quite some time by drifting around the countryside, avoiding identification.

follow up question; was forgery of documents widespread, and were there any common countermeasures used to prevent it? are there any good sources on this that can be used to learn more about this subject?

Just wanted to ask one question What were the rules regarding travelling from one country to another? Could you sneak out to another country and leave behind your criminal life?

http://www.reddit.com/r/AskHistorians/wiki/dailylife#wiki_proof_of_identity

Makes me think of Martin Guerre.