Spycatcher by Peter Wright Page B

an unwise move. Within the year the prison was bombed and many files were destroyed or damaged by fire. Those that could be saved were stored in moisture-resistant polythene bags. In the 1960s, when we began to study the history of recruitments in the 1930s, I often examined prewar files. It was a difficult process, prizing apart the charred pages with tweezers and wooden spatulas.

    After the disaster at Wormwood Scrubs MI5 put a lot of thought into designing an effective Registry. Brigadier Harker, who, as Sir David Petrie's wartime Deputy, was the ideal administrative foil, recruited an expert in business systems, Harold Potter, to reorganize the Registry. Potter was an excellent choice. He had a neat, methodical mind and the will to impose order even in the chaos of wartime.

    In 1955 Potter was approaching retirement, but he took great delight in showing me around. The Registry was based in a central hall, which housed the main file index and the files themselves. The rooms leading off from the central concourse held the other specialist card indexes. Duplicate copies of all files and indexes were routinely made on microfilm, and stored in a specially protected MI5 warehouse in Cheltenham to prevent the catastrophe of Wormwood Scrubs occurring again. Potter's office, tucked in one corner of the Registry, was a paragon of neatness.

    "Make sure you return your files promptly, won't you, Peter? I don't want to have to start chasing you like I do some of these buggers!"

    He could have been a kindly, small-town librarian. Sadly for Potter, I became one of the worst abusers of the Registry, routinely holding scores of files at a time, though never, I suspect, as bad as Millicent Bagot, the legendary old spinster in F Branch who kept tabs on the International Communist Party for decades. I have always assumed Millicent to have been the model for John le Carry's ubiquitous Connie. She was slightly touched, but with an extraordinary memory for facts and files. Potter and his successors in the Registry despaired of Millicent. "I only hope we get the files back when she retires," he would mutter to himself after a particularly heavy file request from F Branch.

    The Registry always fascinated me. Just being there filled me with anticipation, an irresistible feeling that inside the mass of dry paper were warm trails waiting to be followed. Potter explained to me the correct system for signing on and off a file to show that it had been received and dealt with. He had designed the filing system so that each file read chronologically, with papers and attachments on the right, and the index and minutes placed on the left for quick access.

    The whole system depended on accurate and disciplined classification. When an officer wished to file something, it had to be approved by one of Potter's staff. Very often file requests were rejected as being too generalized. When an officer wished to draw a file he filled in a request form. These trace requests were always recorded, and if a trace was requested on an individual more than once, a file was automatically opened on him. There were three basic categories in the Registry. The first category was Personal Files, or PFs, which were buff-colored files arranged in alphabetical order. There were about two million PFs when I joined the Service in 1955. That figure remained fairly static and began to rise dramatically only in the late 1960s and early 1970s with the onset of student and industrial militancy. Then there were subject files, or organizational files, such as for the Communist Party of Great Britain, Subject files very often ran into several volumes and were elaborately cross-referenced with the PFs. The final main category was the duck-egg-blue List File. This generally comprised material gathered during a particular case which could not easily be placed within either of the two previous categories. There were also Y-Boxes. These were a means of separating particularly sensitive files