Errors in the National Firearms Registration and Transfer Record: A New Amnesty Period May be Required to Correct Them by Eric M. Larson Prepared for presentation before the Subcommittee on Treasury, Postal Service and General Government of the Committee on Appropriations House of Representatives 2360 Rayburn House Office Building Washington, D.C. April 8, 1997 TABLE OF CONTENTS Page Consider the Obvious Data Errors that I Pointed Out Three Times. . . . . . . . . . . . . . . . . 1 Most Private Citizens Have Registered NFA Firearms on Form 1 or Form 4467. . . . . . . . . . . . . . . . . . . . . 2 Amnesty Period (1968) and Later Registrations of Unregistered NFA Firearms . . . . . . . . . . . . . . . . . . . 4 Questions About the Accuracy of the NFRTR Have Been Raised, But Have Never Been Resolved. . . . . . . . . . . . . . . . . .10 Some NFRTR Data are Inaccurate Beyond a Reasonable Legal Doubt . . . . . . . . . . . . . . . .13 ATF's certifications of nonregistration are misleading because they give the impression that the NFRTR data are 100 percent accurate. . . . . . . . . . . . . .13 A February 1996 audit of the NFRTR by ATF is flawed because it addressed the wrong postulates. . . . . . . .14 In May 1996, a federal court overturned five convictions for nonregistration because ATF employees deliberately threw registration documents away. . . . . . . . . . . .14 Using Publicly Available Information, Can an Outsider Determine if ATF Has Added a Firearm to the NFRTR After an Owner Produced a Registration? . . . . . . . . . . . .15 If NFRTR Records are Missing, Why Hasn't it Been More Noticeable? . . . . . . . . . . . . . .18 Some NFA firearms are apparently currently registered to persons whom ATF has stated are deceased . . . . . .20 TABLE OF CONTENTS (continued) Page Statistical Probabilities Demonstrate that on Average, the Attempted Transfer of a Registered NFA Firearm for Which ATF May Have Lost the Registration Would be a Relatively Rare Event. . . . . . . . . . . . . . . .21 The Integrity of the NFRTR Data are Questionable Because it Appears that ATF May Have Registered Some Firearms in Violation of the NFA . . . . . . . . . . . . .25 Do any of the 146 Form 4467 registrations that occurred during 1972 to 1986, as reported by ATF in 1996, correspond to firearms that ATF has registered illegally?. . . . . . . . . . . . . . . . . . . . . . . . . . .25 Did ATF illegally register a 28 gauge H&R Handy-Gun bearing serial number 29691 in 1972? . . . . . . . . . .28 Why were five firearms originally manufactured in or before 1934 not registered by the manufacturer until April 16, 1986? . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30 Some Current (1992 to 1996) NFRTR Data are of Questionable Accuracy and Integrity, and the Meaning of Some of These Data is Unclear. . . . . . . .31 The data ATF reports do not appear to reflect changes in total registration that could be expected to occur involving NFA firearms which ATF has removed from the NFA as collector's items. . . . . . . . . . . .34 The significance of increases in the number of registrations for "unknown" years during 1992 to 1996 is unclear, and merits investigation . . . . . . . . . . . . . . . .34 Increases in the overall number of firearms may indicate that ATF personnel have lost or deliberately destroyed original registration documents, and that registrations have been added to the NFRTR whenever a lawful owner was able to produce his or her copy of the registration document. . . . . . . . . . . . . . . . . . . . . . . . . . . .35 TABLE OF CONTENTS (continued) Page Some Current (1992 to 1996) NFRTR Data are of Questionable Accuracy and Integrity, and the Meaning of Some of These Data is Unclear (continued) My direct questions to NFA Branch personnel asking whether ATF has added or would add firearms to the NFRTR after a lawful owner produced a registration or transfer document of which ATF had no record, did not result in any ringing denials . . . . . . . . . . . . . . . . . . . . . . . .36 The basis of certain NFA firearms "REGISTRATION ACTIVITY" particularly in the "OTHER" category, is unclear or unknown . . . . . . . . . . . . . . . . . . . . . . . . . . . .36 ATF personnel identified six types of "Significant Errors" in the NFRTR by June, 19, 1995, and inexplicably downgraded them to "Errors" by October 10, 1995, although all are clearly major errors that affect the accuracy and integrity of the NFRTR . . . . . . . . . . . . . . . . . . . . . . . . . . .37 It appears, based on the preceding discussion and other evidence, that NFRTR data have the potential to be manipulated and may, in fact, already been illegally manipulated . . . .41 It is Clearly Illegal for ATF to Claim That Issuing a Duplication NFA Registration Document to a Lawful Owner is a Matter of ATF's Sole Discretion . . . . . . . . . . . . . . . .43 Why Haven't More People Spoken Out?. . . . . . . . . . . . . . . . . . . . . . .46 Where Do We Go From Here? . . . . . . . . . . . . . . . . . . . . . . . . .48 Endnotes. . . . . . . . . . . . . . . . . . . . . . . . . . . .52 LIST OF TABLES Page Table 1: Year of Original Form 4467 (Amnesty Period) Registrations, as Reported by ATF during 1992 to 1996 in the National Firearms Registration and Transfer Record . . 6 Table 2: Original Form 1 Registrations From 1934 to 1971, and During Unknown Years, as Reported by ATF in the National Firearms Registration and Transfer Record During 1992 to 1996. . . . . . . . . . . . . . . . . . 8 Table 3: Number of Form 1 Registrations from 1934 to 1971, on Form 4467 for all Years, and During "Unknown" Years for Both, that ATF has Apparently Added to the National Firearms Registration and Transfer Record During 1992 to 1996 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Table 4: Form 4 (Tax-Paid) Transfers from 1934 to 1996, and During Unknown Years, as Reported by ATF During 1992 to 1996 in the National Firearms Registration and Transfer Record: Calculations Showing Results of Annual and Overall Changes Have Been Added . . . . . .12 Table 5: Number of NFA Firearms Registered During 1934 to 1971 and Unknown Years: Transferred Since Their Original Registration, and Still Owned Since Their Original Registration Period as Reported by ATF in 1996, by Number and Percent. . .19 Table 6: Numbers of Firearms that ATF Has Apparently Added to the National Firearms Registration and Transfer Record during 1992 to 1996, versus Hypothetical Estimates For These Same Years. . . . . . . . . . . . . . . . . . . . . . . . . . .23 Table 7: Form 4467 (Amnesty Period) Registrations as Reported by ATF in the National Firearms Registration and Transfer Record During 1992 to 1996: Pre-1968 (None are Possible), 1968-71 (Most are Legal) and Post-1971 (All are Illegal). . . . . . . . . . . . . . . . . . . . . . . . . . . .27 Table 8: Total "REGISTRATION ACTIVITY" from 1934 to 1971, and During Unknown Years, as Reported by ATF in the National Firearms Registration and Transfer Record during 1992 to 1996 . . . . . . . . . . . . . . . . . . . . . .33 LIST OF TABLES (continued) Page Table 9: Number of "OTHER" Category "REGISTRATION ACTIVITY" from 1934 to 1996, and During Unknown Years, as Reported by ATF in the National Firearms Registration and Transfer Record During 1992 to 1996. . . . . . . .38 Table 10: Symbols on a Computer Printout of Selected 1995 Data from the National Firearms Registration and Transfer Record, and their Possible Meanings . . . . . . . . . . . . .42 PAGE INSERTS Page Inserts following page. . . . . . . . . . . . . . . . . . . . . 3 1. Approved Form 1 ("Registration of Firearms") dated August 3, 1938, for a .410 H&R Handy-Gun bearing serial number 52003, which had been unregistered and was voluntarily registered by G. L. Daniels of St. Paul, Minnesota 2. Reverse of approved Form 1 registration dated August 3, 1938 3. Affidavit by G. L. Daniels, dated August 3, 1938 4. Notice from Arthur D. Reynolds, Collector, Treasury Department, Office of the Collector of Internal Revenue, St. Paul, Minnesota Insert following page . . . . . . . . . . . . . . . . . . . . . 4 1. Approved Form 1 ("Application for Registration of Firearm") dated December 18, 1961, for a .410 H&R Handy-Gun bearing serial number 13907, which had been unregistered and was was voluntarily registered by Robert Ray Miller of Austin, Texas PAGE INSERTS (continued) Page Inserts following page. . . . . . . . . . . . . . . . . . . . . 5 1. Approved Form 4467 (Registration of Certain Firearms During November 1968") dated January 30, 1969, for a 28 gauge H&R Handy-Gun bearing serial number 7065, which had been unregistered, and was voluntarily registered by Leonard C. Halverson, East Falmouth, Massachusetts 2. Reverse of approved Form 4467 registration dated January 30, 1969, bearing an ATF time/date stamp indicating that the form was received on November 27, 1968 Insert following page . . . . . . . . . . . . . . . . . . . . .11 1. Approved Form 4 ("Application for Tax Paid Transfer and Registration of Firearm") dated October 24, 1989, for a .410 H&R Handy-Gun bearing serial number 50885, which lawfully transferred its ownership from Michael D. Addis of Richfield, Minnesota, to Eric M. Larson of Takoma Park, Maryland 2. Reverse of approved Form 4 transfer and registration dated October 24, 1989 Insert following page . . . . . . . . . . . . . . . . . . . . .25 1. Approved Form 4467 ("Registration of Certain Firearms During November 1968") dated March 14, 1969, by use of an "Approved---Harold A. Serr" stamp, for a caliber 7.62 x 39mm AK-47 Submachine gun bearing serial number 1695, which had been unregistered and was registered by Philip Carl Wahlbom, Jr., of Ft. Wainwright, Arkansas, who submitted it to ATF for registration on March 3, 1969 PAGE INSERTS (continued) Page Insert following page . . . . . . . . . . . . . . . . . . . . .29 1. Approved Form 4 ("Application for Transfer and Registration of Firearm") dated March 23, 1972, for a 28 gauge H&R Handy-Gun bearing serial number 29691, which lawfully transferred its ownership from Alan C. Buck of Houston, Texas, to Robert R. Miller of Houston, Texas, signed by Rex. D. Davis, Director, Bureau of Alcohol, Tobacco and Firearms Inserts following page. . . . . . . . . . . . . . . . . . . . 39 1. Document obtained under a 1996 Freedom of Information Act request: "ERROR RESOLUTION REVIEW" with "Approval Date: Sept. 24, 1993" 2. Document obtained under a 1996 Freedom of Information Act Request "NFA's Monthly PRODUCTIVITY Report" that was "prepared 2/9/95" 3. Document obtained under a 1996 Freedom of Information Act request: National Firearms Act Branch document entitled "WEEKLY Quality Review Report" with notation "Mon,6/19" 4. Document obtained under a 1996 Freedom of Information Act request National Firearms Act Branch document entitled "WEEKLY Quality Review Report" with notation "Mon,10/30" Inserts following page. . . . . . . . . . . . . . . . . . . . .41 1. December 22, 1995, Freedom of Information Act data run, page 30 2. December 22, 1995, Freedom of Information Act data run, page 34 3. December 22, 1995, Freedom of Information Act data run, page 35 4. December 22, 1995, Freedom of Information Act data run, page 44 5. December 22, 1995, Freedom of Information Act data run, page 52 6. December 22, 1995, Freedom of Information Act data run, page 54 7. December 22, 1995, Freedom of Information Act data run, page 57 8. December 22, 1995, Freedom of Information Act data run, page 65 9. December 22, 1995, Freedom of Information Act data run, page 70 10. December 22, 1995, Freedom of Information Act data run, page 73 11. December 22, 1995, Freedom of Information Act data run, page 77 12. December 22, 1995, Freedom of Information Act data run, page 78 13. December 22, 1995, Freedom of Information Act data run, page 81 PAGE INSERTS (continued) Page Insert following page . . . . . . . . . . . . . . . . . . . . .44 1. Excerpt from testimony of Karl T. Frederick, President, National Rifle Association of America, April 18, 1934, before the Committee on Ways and Means, House of Representatives Errors in the National Firearms Registration and Transfer Record: A New Amnesty Period May be Required to Correct Them by Eric M. Larson In 1996, I documented obvious errors in the National Firearms Registration and Transfer Record such as amnesty guns registered before 1968 and that firearms were apparently being added to the NFRTR after lawful owners showed their valid paperwork to agents of the Bureau of Alcohol, Tobacco and Firearms (ATF). Otherwise, why would the number of Form 4467 (amnesty) registrations increase from 57,187 as reported in 1992, to 57,216 as reported in 1995, an increase of 29 guns? In 1971, the U.S. Supreme Court ruled that no more firearms could be registered on Form 4467 that was more than 25 years ago. But at the end of 1996, the number was 57,223 7 more guns than in 1995. That's 36 new gun registrations in just 4 years. What's going on here? Consider the Obvious Data Errors that I Pointed Out Three Times In 1997, these and other errors I documented remain more obvious than ever. Yet, consider when and to whom I pointed these errors out: On April 30, 1996, in testimony before the House Subcommittee on the Treasury, Postal Service, and General Government Appropriations, which funds and oversees ATF; In an article in the July 1996 issue of Machine Gun News, to which ATF headquarters and all of its field offices subscribes; and In a November 21, 1996, letter to ATF Director John W. Magaw. I recently got a response from ATF, but not from Mr. Magaw. Instead, NFA Branch Chief Nereida W. Levine confirmed, in a letter to me dated January 9, 1997, there are errors in the NFRTR as I noted, and stated that "if a person possessed a valid registration that was not in the Registry . . . it would be added." Viewed against the recent NFRTR statistics that ATF produced on January 7, 1997, covering "DATA THROUGH 12/31/96," several of Ms. Levine's statements are particularly striking: In regard to . . . your letter, we are constantly verifying the information in our database. If we do locate a record where the date, form number, or other information was not entered correctly, we enter the correct information. These actions may then result in an adjustment to previously generated statistics. Ms. Levine's comments are more than striking they are stunning: In view of the current NFRTR data, exactly what does "constantly verifying the information in our database" mean? Consider the fact that every year since at least 1992, ATF has reported registrations of NFA firearms during years and in categories which they cannot logically or legally exist. Some registrations are reported in years before 1934, the same year the NFA was enacted. So, how could firearms have been registered during "1920 to 1929" or "prior to 1920"? Also consider pre-1968 Form 4467 registrations. When Congress revised the NFA in 1968, it allowed persons to register unregistered NFA firearms with immunity from prosecution during an amnesty period from November 2, 1968, to December 1, 1968. Therefore, no Form 4467 registration before 1968 can possibly be correct, yet ATF reported there were 164 as of December 31, 1996. Moreover, Form 4467 registrations were prohibited after April 5, 1971, by a U.S. Supreme Court decision. Yet, ATF continues to report post-1971 Form 4467 registrations 146 in 1996. Did ATF register these guns illegally? All of the types of errors I pointed out last year continue to be reported in the most recently published NFRTR data, which are current through December 31, 1996. What's going on here? But first, it is necessary to explain some NFA firearms registration issues that apply to private citizens who own NFA firearms. Most Private Citizens Have Registered NFA Firearms on Form 1 or Form 4467 Relatively few people own NFA firearms. Indeed, the total number of registered NFA firearms as of December 31, 1996, was 838,286, according to ATF. This number excludes most firearms owned by or under the control of the United States Government, such as within the military. This number includes all NFA firearms owned by: (1) private citizens; (2) federally licensed manufacturers, importers and dealers; (3) state or local government entities, including police and sheriff departments, museums, and prisons; and (4) private companies, such as those offering security protection. If one believes that there are at least 250 million firearms of all types in the United States today, then 838,286 represents about one-third of 1 percent (that is, 0.33 of 1 percent) of all 250 million firearms. To the best of my knowledge no definitive count has recently been done, but I believe that the vast majority of registered NFA firearms probably are not owned by private citizens. Also, many registered NFA firearms are exported. Nearly all the of the original registrations of NFA firearms by private citizens, which were allowed under the NFA from 1934 to 1971, were done using Form 1 or Form 4467, as explained below. Transfers of registered NFA firearms to private citizens who are not federally licensed NFA firearm dealers are usually done on Form 4 (tax-paid transfer) or, less typically, on Form 5 (tax-free transfer, such as through inheritance). Various other forms used to manufacture, transfer, import or export NFA firearms will not be discussed here in detail because they are generally used only by federally licensed manufacturers, importers, exporters or dealers in NFA firearms. All NFA firearm registrations by and transfers to private individuals are made in duplicate original one copy for the lawful owner, and the other copy for ATF. The person registering the firearm, or the transferee, must personally sign each copy of the document. Under the original NFA, which became effective on July 26, 1934, any person who came into possession of an unregistered NFA firearm was under a continuing obligation to register it, although there was considerable discretion involving the application of a penalty for failure to do so. In fact, under the original NFA and until September 1, 1952, the law provided that a person had up to 60 days to voluntarily register an unregistered NFA firearm that came into his or her possession without penalty but depending upon the circumstances, a penalty may not have applied even if the firearm was registered later than that. During July 26, 1934, to October 31, 1968, most private citizens who voluntarily registered unregistered NFA firearms did so using Form 1. According to the most recent (1996) statistics, more than 32,000 NFA firearms were registered on Form 1 during the 34-year period from 1934 to 1968. In general, since November 1, 1968, Form 1 has been used to apply for permission from ATF to "make" and register an NFA firearm. Private citizens voluntarily (and routinely) registered NFA firearms on Form 1 for many years without penalty. An example of an approved Form 1 that was used to register an unregistered .410 H&R Handy-Gun dated August 3, 1938, has been inserted following this page. Also inserted are copies of (1) the reverse of the Form 1, containing instructions; (2) an Affidavit by the registrant attesting to the facts of the registration; and (3) a notice from the Treasury Department, Office of the Collector of Internal Revenue, St. Paul, Minnesota, where the firearm was accepted for registration. Under current law, NFA firearms are no longer registered with the local collectors of internal revenue. Insert pages here All of the NFA forms changed slightly through the years. An example of a Form 1 that was approved 23 years later, on December 18, 1961, also to register an unregistered .410 H&R Handy-Gun, differs little in content. One difference is that in 1961, the affidavit procedure was replaced by a simple declaration, and approved locally. Note that this Form 1 registration document was stamped nonwillful violation approval recommended. This meant that while the person who possessed the firearm had violated the law, the law also was sufficiently flexible at that time to allow a person to register an unregistered NFA firearm without penalty. A copy of this Form 1 has been inserted following this page. Insert page here When the Congress extensively revised the NFA under Title II of the Gun Control Act (GCA) of 1968 it did so, in part, to remedy a constitutional defect in the original NFA. In January 1968, the U.S. Supreme Court ruled that the NFA's registration provision was unconstitutional, because it required persons to register NFA firearms and ATF to make these data available to local, state, and other federal officials upon request. But persons who possessed NFA firearms in violation of state or local law(s) risked the hazards of prosecution by supplying the registration information required by ATF, which violated their 5th Amendment rights, guaranteed by the U.S. Constitution, against self-incrimination. The Congress resolved this conflict in amending the NFA under Title II of the 1968 Act by: (1) prohibiting any information required to comply with the NFA to be used against a registrant or applicant "in a criminal proceeding with respect to a violation of law occurring prior to or concurrently with the filing of the application or registration, or the compiling of the records containing the information or evidence"; (2) establishing an amnesty period from November 2, 1968, to December 1, 1968, when persons could register unregistered NFA firearms with full immunity from prosecution; and (3) prohibiting the release of any information about the registration status or ownership of any NFA firearm. The 1968 amnesty period is the first and only one the Government has ever held. So-called "amnesty" guns were registered from November 2, 1968, to December 1, 1968, on Form 4467 (entitled "Registration of Certain Firearms During November 1968"). Amnesty Period (1968) and Later Registrations of Unregistered NFA Firearms The legislative history of the amnesty provision states the Congress intended that "every firearm in the United States should be registered to the person possessing the firearm" by December 2, 1968, the day after the 30-day amnesty period expired. Except for providing in law that the Secretary of the Treasury may authorize future amnesty periods as needed to "contribute to the purposes" of the NFA, Congressional intent regarding the voluntary registration of unregistered NFA firearms after the amnesty period expired is not specifically discussed. On Form 4467, the following statement appears above data block 1: IMPORTANT This form cannot be accepted for registration of firearm except when received by Director during the time period November 2, 1968, through December 1, 1968. An example of an approved Form 4467 that was used to register an unregistered 28 gauge H&R Handy-Gun dated January 30, 1969, has been inserted following this page. Also inserted is a copy of the reverse of this Form 4467, which contains instructions and bears a Treasury Department date stamp indicating that the form was received November 27, 1968. Insert pages here Gary N. Schaible, a firearms specialist in the NFA Branch, told me during an April 1996 interview that ATF continued to accept Form 4467 registrations after December 1, 1968, from U.S. military service personnel who returned to the United States after completing an overseas tour of duty. The reason, he said, was that these personnel had not been in the United States and were thus unable to participate in the 1968 amnesty. While it is not commonly known, ATF also continued to allow private citizens to register unregistered NFA firearms by approving a Form 4 transfer, using the transfer as a method to register a previously unregistered firearm to the transferee. According to ATF Chief Counsel Jack B. Patterson, ATF accepted and approved Form 4 applications to transfer and register unregistered NFA firearms by individuals until the U.S. Supreme Court prohibited such registrations on April 5, 1971. Unregistered firearms could also be registered on Form 10 by a federal, state, or local government entity, and subsequently transferred to individuals. ATF itself recommended this "pass through" practice to persons who came into possession of unregistered NFA firearms and wanted to register them. In July 1973, however, ATF prohibited (by regulation) the transfer of any firearm registered on Form 10 other than to a government entity. While unregistered NFA firearms can still be registered by government entities on Form 10, no firearm that is registered on Form 10 can be transferred to individuals, federally licensed firearms dealers, or federally licensed gun collectors regardless of the date or conditions under which it was actually registered. Thus, under current law, no person is allowed to voluntarily register an unregistered NFA firearm, or to re-register the firearm if ATF loses the registration record. Table 1, which lists Form 4467 registrations as reported by ATF for each year from 1992 to 1996, shows 146 registrations after 1971 for the year 1996. Yet, Form 4467 registrations were halted by court order in 1971 more than 25 years ago. There were 172 such registrations in 1995, and the reduction to 146 apparently involves reclassification. But also note that the total number of Form 4467 registrations increased from 57,216 in 1995 to 57,223 in 1996 that is, 7 guns were added. Where did these guns come from? Table 1 about here Table 1 shows other logical impossibilities. The number of Form 4467 registrations before 1968 that ATF has reported in 1996 was 164. No Form 4467 registration for any year before 1968 can possibly be correct. Table 1 Year of Original Form 4467 (Amnesty Period) Registrations, as Reported by ATF during 1992 to 1996 in the National Firearms Registration and Transfer Record Year of Net change in Net number original registrations, added or Form 4467 Years of registration activity 1992-1996 removed, registration 1992 1993 1994 1995 1996 (6) - (2) = 1992-1996 (1) (2) (3) (4) (5) (6) (7) (8) 1988 1 0 1987. . . . 1 1 1 1 -1 -1 1986. . . . 3 3 3 3 1 -2 -2 1985. . . . 1 1 1 1 -1 -1 1984. . . . 3 3 3 3 -3 -3 1983. . . . 1 1 1 1 -1 -1 1982. . . . 3 3 3 3 -3 -3 1981. . . . 1 1 1 1 -1 -1 1980. . . . 4 4 4 4 -4 -4 1979. . . . 1 1 1 1 -1 -1 1978. . . . 6 6 6 6 -6 -6 1977. . . . 1 1 1 1 1 0 1976. . . .39 39 39 39 36 -3 -3 1975 1974. . . . 5 5 5 5 5 0 1973. . . .17 17 18 18 18 1 +1 1972. . . .84 84 84 84 85 1 +1 1971. . . .26 26 26 26 26 0 1970. . . 272 272 271 271 271 -1 -1 1969. . 2,006 2,010 2,016 2,017 2,053 47 +47 1968. .54,48754,485 54,485 54,503 54,505 18 +18 1967. . . .64 64 64 64 64 0 1966. . . . 8 8 8 8 8 0 1965. . . . 2 2 2 2 2 0 1964. . . . 3 3 3 3 3 0 1963. . . . 2 2 2 2 2 0 1962. . . .14 14 14 14 14 0 1961. . . . 2 2 2 2 2 0 1960. . . . 6 6 6 6 6 0 1950-59 . .23 23 23 23 23 0 1940-49 . . 9 9 9 9 10 1 +1 1930-39 . .25 25 25 25 27 2 +2 1920-29 . . 6 6 6 6 3 -3 -3 before 1920 . . . 4 4 4 4 -4 -4 Unknown . .58 58 57 57 58 0 Totals.57,18757,189 57,196 57,216 57,223 36 +36 Change from n/a +2 +7 +20 +7 36 +36 previous year Data Source: Bureau of Alcohol, Tobacco and Firearms. All numbers shown in boldface type were calculated by Eric M. Larson. Blank spaces represent zero. Note: The Congress authorized a 30-day amnesty period only from November 2, 1968, to December 1, 1968. Therefore, all Form 4467 registrations listed for 1967 and earlier are errors, and all Form 4467 registrations that ATF accepted after December 1, 1968, are of questionable legality. Similar patterns occur in Form 1 registrations from 1934 to 1971. As Table 2 shows, 83 firearms were added from 1992 to 1996 for registrations in or before 1971, including 48 registrations during "unknown" year(s). Registrations for "unknown" years increased from 12 to 30 during 1993 to 1994, and from 30 to 60 during 1994 to 1995. Why weren't these registrations classified in the years they were entered? If ATF originally received (or re-recorded) some of the registrations listed in Table 2 during 1992 to 1996, why aren't they listed in the appropriate year(s)? The years of registration (1994 and 1995) that significant numbers of "unknown" year guns were added to the NFRTR are obviously known to ATF. Where did these "unknown" registrations turn up? Is that information known? Are these errors, or are firearms being added? Table 2 about here The results of my calculations in Table 1 and Table 2 are summarized in Table 3. The data in Table 3 represent the number of NFA firearm transfers completed for which ATF may have had no record of the original registration but the lawful owner did. ATF would have thus been forced to add the firearm to the NFRTR. Table 3 about here The data in Table 3 are based upon the actual NFA firearm registration data that ATF reported for each year during 1992 to 1996, for (1) registrations on Form 1 from 1934 to 1971, including "unknown" years; and (2) all Form 4467 registrations, including those for "unknown" years. While having calculated the data shown in Table 3 from actual NFRTR data, I would nevertheless advise readers to be somewhat cautious in drawing exact conclusions from it. The reason is that these data could underestimate the number of firearms ATF may have added to the NFRTR, for at least two reasons: 1. The data in Table 3 may not include firearms which ATF has seized because the original registrant couldn't produce his or her paperwork. Consequently, to the extent (if any) that these actual data represent firearms ATF has added to the NFRTR, the data in Table 3 would underestimate the number of registrations that ATF may have lost or destroyed and for which the lawful owner has lost the registration. 2. The data in Table 3 may not reflect the effect of ATF having removed firearms from the NFA as collector's items. As we have seen, during each year from 1992 to 1996 there has been a net increase in the total number of certain Form 1, and all Form 4467, firearm registrations. Thus, to the extent that ATF has removed any firearm originally registered on Form 1 during 1934 to 1971, on Form 4467 during any year (or during any "unknown" year), the data in Table 3 would clearly underestimate the actual number of firearms added during 1992 to 1996. Table 2 Original Form 1 Registrations from 1934 to 1971, and During Unknown Years, as Reported by ATF in the National Firearms Registration and Transfer Record During 1992 to 1996 Net number Year of Net change of firearms original in Form 1 added or Form 1 Years of registration activity registrations, removed, registration 1992 1993 1994 1995 1996 1992-1996 1992-1996 (6) - (2) = (1) (2) (3) (4) (5) (6) (7) (8) 1971. . . 24 24 24 24 24 0 1970. . . 38 38 38 38 39 1 +1 1969. . . 36 36 36 36 36 0 1968. .1,510 1,509 1,510 1,510 1,511 1 +1 1967. . .909 909 909 909 909 0 1966. . .900 900 902 902 902 2 +2 1965. . .841 841 841 843 843 2 +2 1964. . .744 744 744 744 744 0 1963. . .709 709 709 709 709 0 1962. . .734 734 734 734 734 0 1961. . .809 810 810 811 812 3 +3 1960. . .790 791 792 792 792 2 +2 1950-596,629 6,629 6,631 6,633 6,638 9 +9 1940-496,574 6,572 6,571 6,574 6,575 1 +1 1930-3911,422 11,422 11,422 11,427 11,449 27 +27 1920-29 . 12 12 12 12 -12 -12 before 1920 . . . .1 1 1 1 -1 -1 Unknown . 12 12 30 60 60 48 +48 Totals32,694 32,693 32,716 32,759 32,777 83 +83 Change from n/a -1 +23 +43 +18 83 +83 previous year Data Source: Bureau of Alcohol, Tobacco and Firearms. All numbers shown in boldface type were calculated by Eric M. Larson. Blank spaces represent zero. Note: The National Firearms Act (NFA) was enacted in 1934, and became effective on July 26 of that year. All registrations ATF has reported for years before 1934 are errors. Table 3 Number of Form 1 Registrations from 1934 to 1971, on Form 4467 for all Years, and During "Unknown" Years for Both, that ATF has Apparently Added to the National Firearms Registration and Transfer Record During 1992 to 1996 Registration type 1993 1994 1995 1996 TotalsAnnual averages (1)+(4)=(5) 4 years = (1) (2) (3) (4) (5) (6) Form 1. . . . . .2 7 20 7. . . 36. . . 9.00 per year Form 4467 . . . -1 23 43 18. . . 83. . .20.75 per year Totals . . . .1 30 63 25. . . . . . . . 119 29.75 per year Data source: Table 1 and Table 2 of this article; the year 1992 is not listed here because it is the base year from which these comparisons originate. The reason is that if the registrations of firearms removed from the NFA are removed from the NFRTR data base, by definition this must reduce the total number of firearms registered. Thus, to the extent that ATF may have removed firearms, any increases in the total number must be larger than the total shown for that year by that same amount. Overall, 119 firearms of questionable origin were added to the NFRTR within certain categories during 1992 to 1996, as Table 3 shows. These consist of 36 Form 4467 registrations (which were supposed to have ended in 1971) and 83 Form 1 registrations in or before 1968 (including 48 for "unknown" years). Do these 119 firearms an average of nearly 30 firearms per year represent reclassification of registered firearms, firearms whose registrations were "lost and found" and then added to the NFRTR by ATF, or both or is there some other explanation? The answer to this question is of more than just academic interest it is of intense legal interest. The reason is that if a legally valid registration document is found to be missing from the NFRTR, the Department of Justice has determined "the only solution would be to declare another amnesty period." Under  207(d) of the Gun Control Act of 1968: The Secretary of the Treasury, after publication in the Federal Register of his intention to do so, is authorized to establish such periods of amnesty, not to exceed ninety days in the case of any single period, and immunity from liability during any such period, as the Secretary determines will contribute to the purposes of this title. Because the legislative history of  207(d) is not more specific, it is difficult to be certain of the conditions under which the Congress intended amnesty periods to be authorized. I believe that such a condition would include that of ATF losing original firearm registration documents. The Department of Justice agrees. Questions About the Accuracy of the NFRTR Have Been Raised, But Have Never Been Resolved In 1975, an internal ATF study of the NFRTR documented the following problem: "Records missing as evidenced by registered owner producing registration form when no form or index card was located in the NFA." In 1979, an internal ATF memorandum commenting upon the 1975 study stated that "To our knowledge, this has never occurred." "If both the index card and transaction record are missing," the 1979 memorandum says, "it would represent a complete breakdown of the system." On October 18, 1995, Thomas Busey, then Chief of the National Firearms Act Branch, made several remarks about the accuracy of the NFRTR via closed-circuit television at ATF headquarters, in a "Roll Call Training" session. He said, in part: Let me say that when we testify in court, we testify that the data base is 100 percent accurate. That's what we testify to, and we will always testify to that. As you probably well know, that may not be 100 percent true . . . we're hoping [that numerous cross-checks using multiple identifiers] eliminates the possibility that anything goes out erroneous because we know you're basing your warrants on it, you're basing your entries on it, and you certainly don't want a Form 4 waved in your face when you go in there to show that the guy does have a legally-registered Title 2 weapon. I've heard that's happened. I'm not sure . . . when I first came in a year ago, our error rate was between 49 and 50 percent, so you can imagine what the accuracy of the [NFRTR] could be, if your error rate's 49 to 50 percent. Mr. Busey's statement is disturbing for at least two reasons. First, he stated that ATF personnel routinely commit perjury, because they testify that the NFRTR is 100 percent accurate and know that it is not. Second, Mr. Busey expresses concerns "that anything goes out erroneous." The perjury aspect will be discussed in a later section of this study. For now, let's see if there's any evidence in the NFRTR data that applies to "a Form 4 waved in your face." A Form 4 (tax-paid) transfer, as we have seen, is most likely to be used by a private citizen. An example of both sides of an approved Form 4 dated October 24, 1989, has been inserted following this page. Insert pages here Table 4 about here Table 4 lists Form 4 transfers from 1992 to 1996 by year, as reported by ATF, which are the most recent data available. Simple inspection of these data shows that in 1995 the year Mr. Busey made his remark 100 Form 1 documents were added or reclassified since 1994. In 1996, the number of added or reclassified Form 1s increased to 283; and of these 283 Form 4 documents, 192 (almost 68 percent) occurred for transfers before 1968 nearly 30 years ago. What is going on here? At the least, it appears that there is plenty of evidence to support Mr. Busey's concerns. In fact, the data in Table 4 clearly show it is possible that ATF agents had Form 4s "waved in their faces" as many as 625 times during 1992 to 1996. What Mr. Busey reported i.e. "a Form 4 waved in your face when you go in there to show that the guy does have a legally-registered Title 2 weapon" isn't new information to some people. A lack of up-to-date record-keeping undoubtedly was the reason that during a 1977 compliance audit of J. Curtis Earl of Phoenix, Arizona, ATF apparently had no record of 425 NFA firearms which Mr. Earl ultimately proved were legally registered to his business. Table 4 Form 4 (Tax-Paid) Transfers from 1934 to 1996, and During Unknown Years, as Reported by ATF During 1992 to 1996 in the National Firearms Registration and Transfer Record: Calculations Showing Results of Annual and Overall Changes Have Been Added Change, Year 1992 1993 Change 1994 Change 1995 Change 1996 Change 1992-96 (2)-(1)= (4)-(2)= (6)- (4)= (8)-(6)= (1) (2) (3) (4) (5) (6) (7) (8) (9) 1996 6,367 00 1995 8,059 0 8,086 +27+27 1994 7,838 07,870 +32 7,887 +17+49 1993 7,749 0 7,819 +707,837 +18 7,850 +13+101 1992 6,527 6,556 +29 6,568 +126,573 +5 6,577 +4+50 1991 5,390 5,400 +10 5,411 +115,420 +9 5,423 +3+33 1990 6,807 6,821 +14 6,830 +96,835 +5 6,841 +6+34 1989 8,165 8,176 +11 8,176 08,181 +5 8,186 +5+21 1988 7,699 7,703 +4 7,707 +47,712 +5 7,714 +2+15 1987 8,311 8,318 +7 8,321 +38,330 +9 8,331 +1+20 1986 5,158 5,162 +4 5,172 +105,174 +2 5,174 0+16 1985 3,524 3,526 +2 3,529 +33,532 +3 3,537 +5+13 1984 3,911 3,913 +2 3,915 +23,916 +1 3,919 +3+8 1983 3,203 3,204 +1 3,207 +33,207 0 3,208 +1+5 1982 2,770 2,771 +1 2,770 -12,770 0 2,771 +1+1 1981 3,734 3,735 +1 3,737 +23,741 +4 3,741 0+7 1980 3,040 3,040 0 3,044 +43,046 +2 3,046 0+6 1979 2,150 2,150 0 2,151 +12,151 0 2,151 0+1 1978 1,879 1,878 -1 1,879 +11,878 -1 1,878 0-1 1977 1,535 1,535 0 1,537 +21,537 0 1,538 +1+3 1976 979 979 0 983 +4 983 0 983 0+4 1975 567 567 0 567 0 568 +1 569 +1+2 1974 579 579 0 579 0 579 0 579 00 1973 353 353 0 353 0 353 0 354 +1+1 1972 261 261 0 261 0 262 +1 262 0+1 1971 36 36 0 36 0 36 0 36 00 1970 10 10 0 10 0 10 0 10 00 1969 13 13 0 13 0 13 0 13 00 1968 192 193 +1 193 0 194 +1 194 0+2 < 1968 2,780 2,785 +5 2,792 +72,791 -1 2,983 +192+203 Unknown 22 23 +1 26 +3 25 -1 25 0+3 CHANGE +92 +150 +100 +283 +625 Totals 79,573 87,413 95,398 103,558 110,014 Data source: Bureau of Alcohol, Tobacco and Firearms. All numbers shown in boldface type were calculated by Eric M. Larson. Consider this situation: A legally registered NFA firearm for which ATF has no record? Now think about the plain meaning of Mr. Busey's words as they regard Form 4s: "I've heard that's happened." To see if that information could be confirmed by anybody else at ATF, in April 1996 I asked Gary N. Schaible, the custodian of NFA records (Mr. Schaible's job title is Firearms Specialist in the NFA Branch), if Mr. Busey's statement was correct, and that ATF had added firearms to the NFRTR after the lawful owner showed ATF his copy of the registration document. Mr. Schaible told me: "Yes. I assume that's happened." I asked him again, to make sure I correctly understood his response. Again, Mr. Schaible said: "Yes. I assume that's happened." Does this "represent a complete breakdown of the system?" From the plain meaning of Mr. Busey's and Mr. Schaible's statements, and the evidence presented in Tables 1 to 4, it is difficult to avoid that conclusion. The Department of Justice has taken the Busey matter very seriously. On May 20, 1996, Frederick W. Thieman, Esq., United States Attorney, and Mary Beth Buchanan, Esq., Assistant U.S. Attorney, of the office of the United States Attorney, Western District of Pennsylvania, mailed a thick packet of materials to attorneys representing every defendant in the United States charged with nonregistration of NFA firearms. These materials were sent out, they stated, "in order to avoid any suggestion that the Department [of Justice] has not provided all material which may be relevant in this matter." The packet contains a transcript of Mr. Busey's remarks on October 18, 1995, and "other documents that relate to Mr. Busey's statements." The materials include six "Packet Exhibits" pertaining to the accuracy of the NFRTR. Some NFRTR Data are Inaccurate Beyond a Reasonable Legal Doubt Let's examine three aspects of nonregistration issues: (1) ATF's statements about what its certifications of nonregistration represent; (2) ATF's February 1996 in- house audit of the accuracy of the NFRTR database; and (3) a May 1996 federal court decision that overturned convictions for nonregistration because ATF employees deliberately threw registration documents away. As I will demonstrate, some NFRTR data are inaccurate beyond a reasonable legal doubt. 1. ATF's certifications of nonregistration are misleading because they give the impression that the NFRTR data are 100 percent accurate. ATF's statements about what its certifications of nonregistration represent are too narrowly expressed, as they apply to accuracy of the NFRTR data. Consider that ATF has testified before the Congress that "no one has ever produced a registration form or testified that he had registered a particular weapon after ATF had certified to a court that the person was not the registered possessor of a particular weapon." That's simply not the whole story. According to Robert E. Sanders, Esq., a former head of ATF's Criminal Investigations Division which is ATF's law enforcement function, and is today called Criminal Enforcement Programs that testimony is at least partially misleading. During a September 1996 interview, Mr. Sanders told me: Because of the way these things are handled internally at ATF, the case of a person who was accused of possessing an unregistered firearm who ultimately produced a valid registration or transfer document would never reach the certification stage. ATF would simply add the registration or transfer data to its records, and the case would be dropped. It would not, in other words, make sense for ATF to certify that a firearm for which a registration document was located was not registered. "That [certification of nonregistration] would never happen," Mr. Sanders added. "It would be settled long before the case reached a courtroom." 2. A February 1996 audit of the NFRTR by ATF is flawed because it addressed the wrong postulates. An internal ATF audit of the NFRTR in February 1996 in response to Mr. Busey's statements is misleading, because it addresses the wrong postulates. This audit tested "if an NFA name search would find the name being searched and then if the search would find supporting registration documents" and found that it "conclusively" supported "the reliability of name searches in the database, and related searches for supporting registration documents." The postulates ATF used in this instance are immaterial because it is obvious that no record search, however, diligent, can locate a record of a valid NFA registration document that is not in the NFRTR. They also do not address the question of whether a validly registered firearm was added to the NFRTR, for which ATF had lost or destroyed the record. Neither do they address the question of whether ATF personnel may have illegally registered NFA firearms. Presumably, as will be discussed later in this article, these postulates could be answered conclusively by an independent forensic audit of the NFRTR. 3. In May 1996, a federal court overturned five convictions for nonregistration because ATF employees deliberately threw registration documents away. On May 21, 1996, NFA Branch firearms specialist Gary N. Schaible testified, in a court case appealing convictions for nonregistration, that some ATF employees have thrown registration or transfer documents away because they didn't feel like doing the work to enter them into the NFRTR. The United States Attorney prosecuting the case declined to cross-examine Mr. Schaible, whose testimony was good enough for U.S. District Judge John A. MacKenzie. "I'm going to throw out the convictions that have to do with registrations," he stated, adding that "it throws a disagreeable proposition on my finding somebody guilty on records when their chief man says they were 49 percent wrong." Mr. Schaible's testimony, and especially the court decision, casts reasonable legal doubt on the validity and reliability of NFRTR data. ATF has not appealed the dismissals. Classification errors in reporting statistical data, and even errors of ATF registering or transferring an NFA firearm contrary to law, are not as legally troublesome to the owner of an NFA firearm as the possibility of ATF losing (or throwing away) his original registration document. The reason is that possession of an unregistered NFA firearm invokes serious criminal penalties: a felony conviction, up to 10 years in jail and up to a $250,000 fine. Using Publicly Available Information, Can an Outsider Determine if ATF Has Added a Firearm to the NFRTR After an Owner Produced a Registration? Using publicly available information, can an outsider determine whether ATF has added registered firearms to the NFRTR, after an owner produced a valid Form 4467 or Form 1 registration document? The answer appears to be "possibly, yes." How to prove it may rest with data in the NFRTR specifically, the date the document (identified by its "docket number" or similar identifier) was received by ATF and the date of original registration. This information is, according to ATF, not protected from public disclosure. A 1980 internal ATF memorandum discusses "a docket book maintained by the NFA Branch in which all incoming transfer forms are recorded." The memorandum advises the ATF Director on Freedom of Information Act disclosure policy, and states that the docket book contains Type of form received, date of form, date of receipt by ATF, names of the transferor and transferee, type of action taken by ATF, and date ATF mailed out notification of its action. Since the [docket] number does not identify a particular taxpayer, it falls outside the definition of return information in section 6103(b)(2). . . . Also, the number does not fit within the definition of "taxpayer identity" under section 6103(b)(6). Finally, the NFA Branch could not determine any demonstrable harm in disclosing this number that would justify exercising the internal practices exemption under 5 U.S.C. Section 552(b)(2). Therefore, we recommend disclosure of docket numbers. In 1996, ATF agreed to release under an FOIA request the date that a firearm was originally registered by ATF to the current lawful owner of the firearm. This might be a crack in the dike, because ATF has determined that the original registration date is not protected information to a lawful owner who obviously knows the firearm's serial number. Consequently, it is doubtful that the release of large numbers of original registration dates that are not associated with any particular firearms or persons could be considered protected information. Thus, the dates that ATF originally registered firearms for particular years (such as Form 4467 registrations in 1968, or Form 1 registrations for each of the years from 1934 to 1971) would not identify any individual taxpayers. Indeed, on page 6 of the 1980 Memorandum, ATF states that information which "does not include data in a form which cannot be associated with, or otherwise identify, directly or indirectly, a particular taxpayer" is not exempt from disclosure. Could the release of a date of original registration paired with its corresponding docket number somehow identify "a particular taxpayer"? I doubt it. I don't see how. Not without a taxpayer's name, or the serial number of the firearm. An FOIA request for ATF to release the dates of original registration for large groups of firearms, by year, along with their corresponding "docket numbers," could provide the information necessary to determine if ATF has been adding firearms to the NFRTR. The idea is that if ATF was confronted in 1992 with, say, a valid registration from 1937 for which ATF had no record, ATF would add the firearm to the NFRTR, as NFA Branch Chief Nereida W. Levine advised me in her January 9, 1997, letter. The original dates on the registration or transfer could be entered into the computer with nobody otherwise the wiser a printout would show the actual dates listed on the document. However, in this instance, the "docket number" (or similar identifier of the individual transaction) would be from 1992. It obviously would not correspond to a "docket number" for a much earlier period. Nor would it be in the docket number range established for most firearms registered during that much earlier year or period unless the docket number had been re-used. Right now I don't know precisely how NFRTR data are currently entered into its data base or if the "docket number" identified in the 1980 ATF memorandum is a part of the data record field. I am sure that the "docket number" or its equivalent exists somewhere if not within the computerized NFRTR, then in a backup system. There has to be a docket number or some equivalent unique identifier in the computerized NFRTR, however. The reason is that ATF otherwise couldn't run the system and produce the data crosstabulations it has published for the past 10 or so years. Each data record has to have a unique identifier, even if part or all of the record is linkable with other records (by serial number or owner's name, for example) and other data bases. Defining terms and understanding how ATF uses them is critical, which is why access to the NFRTR data entry manuals is essential. For example, I purchased the 28 gauge H&R Handy-Gun bearing serial number 7065 that was registered on the Form 4467 inserted following page 5 in this study. The owner who registered the gun during the 1968 amnesty period gave me his copy of the Form 4467, which indicates he signed it to register the gun on November 25, 1968 (see inserts following page 5 of this study). A stamped date in Block 20 on the form indicates that ATF registered the firearm in the NFRTR on January 30, 1969. An ATF date/time stamp on the back of the form indicates that ATF received it on November 27, 1968. In my recent FOIA request to obtain the dates of original registration for various NFA firearms that I own, ATF stated that the date was November 25, 1968. Whether the "original date of registration" is the date the applicant signed the form, the date ATF received the form, or the date that ATF formally approved the registration, or some other date, is good example of a potentially critical analytical detail. I had requested the date that ATF registered the firearm not the date of the application by the registrant, which is what ATF sent me. Moreover, ATF stated that the information it provided is "the original dates that your NFA firearms were registered with Bureau of Alcohol, Tobacco and Firearms." This example shows that without access to the original paperwork, an uninformed interpretation of ATF's response to a question about the NFRTR data could have been seriously misleading although in this instance it was not. It is also valid and reliable evidence that ATF considers the date that the registrant signed the application to be the original registration date, and that the "docket number" or similar identifier, matched with the corresponding date that ATF logged the transaction into the NFRTR or other data base, may be a way to detect "lost and found" NFA firearm registration documents. This methodology could also be used to analyze specific changes in the NFA data, such as I've pointed out in Tables 1 and 2 in this article, which indicate that ATF may be adding firearms to the NFRTR. But a comprehensive evaluation of the NFRTR should include more data (such as from Form 3, Form 4, and other registration or transfer forms), rather than the few examples I've identified. Specifically, ATF could be asked to disclose the date of original registration and the corresponding "docket number" (or similar identifier of the transaction) for each individual firearm that appears to have been added to the NFRTR. If the docket number is in the same range as those for the year in question (say, 1992, when the firearm appears to have been added), that's a pretty strong evidence that ATF has added the firearm in 1992 but claimed an original registration date of 1968 (or whatever year it actually was). The case of "adding" a missing transfer (versus a registration) document would show up differently, and probably require a printout that crosstabulated ranges of docket numbers with individual years of transfers. Consider a hypothetical situation where a firearm was originally registered in 1938 and its ownership was legally transferred on a Form 4 to another person in 1950, but ATF lost or destroyed its copy of that 1950 transfer, or record of it. Suppose that the 1950 transferee decides to sell (transfer) the firearm in 1993? This would be the sequence of events: ATF disapproves the sale (transfer) because the firearm shows up as registered to the original 1938 registrant, and moves to seize it; the owner waves the Form 4 transfer approved in 1950 in ATF's face; then, ATF adds the record of approved transfer to the NFRTR and completes the transfer. In this case, the transfer would be listed in the NFRTR as (legally, and correctly) having been completed in 1950; however, the transaction (transfer) data record would by definition probably have a 1993 docket number. If NFRTR Records are Missing, Why Hasn't it Been More Noticeable? If valid registrations are missing from the NFRTR, why isn't it more noticeable? The answer is that certain NFA firearms are sold or transferred relatively infrequently, especially those registered on Form 1 during 1934 to 1939, and Form 4467 registrations in 1968. As we have seen, these are the forms that have been used by most private citizens to originally register NFA firearms. Consequently, encountering a registered firearm for which there is no record in the NFRTR would be by definition an even rarer event, as I demonstrate below. The data ATF has publicly released so far do not allow precise calculations to be made for particular combinations of registrations and transfers by form number and year. Nevertheless, it is still possible to show that the legal transfer of certain NFA firearms that I have discussed in this article is a relatively rare event, as shown in Table 5. Table 5 about here In Table 5, I have listed the total number of NFA firearms ATF reports were registered during 1934 to 1971, transferred since their original registration period, and still owned since their original registration period, by number and percent, including those for "unknown" years. These data are taken from ATF's 1996 NFRTR data tables entitled "NFA REGISTRATION ACTIVITY ANNUAL COMPARISON" and "NFA WEAPONS INVENTORY BY YEAR OF CURRENT OWNERSHIP." The "REGISTRATION ACTIVITY" data table lists numbers of original NFA firearm registrations by year or group of years. For example, there were 14,259 registrations during "1930 to 1939" (it should be 1934 to 1939, but I'm quoting ATF, and hereafter I'll use 1934). The "YEAR OF CURRENT OWNERSHIP" data table lists the NFA firearms inventory for these same registration periods. This means, for example, that of the 14,259 firearms registered during 1934 to 1939, 11,175 of those are currently owned by persons or institutions that obtained them by registering them, or receiving them by lawful transfer, during that same period. In other words, just 3,084 (fewer than 22 percent) of this original group of 14,259 NFA firearms have been lawfully transferred from 1940 to 1996. Table 5 Number of NFA Firearms Registered During 1934 to 1971 and Unknown Years, Transferred Since Their Original Registration, and Still Owned Since Their Original Registration Period as Reported by ATF in 1996, by Number and Percent Firearms transferred Firearms still owned Year firearms were Number of firearms since their original since their original originally registered originally registered registration period registration period (1) = (2) +(3) 1971. . . . . . . . 5,095 (100%). . . 3,405 (66.8%)1,690(33.2%) 1970. . . . . . . . 2,167 (100%). . . 1,207 (55.7%)960(44.3%) 1969. . . . . . . . 4,114 (100%). . . 1,702 (41.4%)2,412(58.6%) 1968. . . . . . . .58,904 (100%). . . 8,590 (14.6%)50,314(85.4%) 1967. . . . . . . . 3,461 (100%). . . 1,913 (55.3%)1,548(44.7%) 1966. . . . . . . . 3,858 (100%). . . 2,123 (55.0%)1,735(45.0%) 1965. . . . . . . . 3,739 (100%). . . 1,979 (52.9%)1,760(47.1%) 1964. . . . . . . . 2,804 (100%). . . 1,553 (55.4%)1,251(44.6%) 1963. . . . . . . . 2,672 (100%). . . 1,333 (49.9%)1,339(50.1%) 1962. . . . . . . . 3,143 (100%). . . 1,703 (54.2%)1,440(45.8%) 1961. . . . . . . . 4,323 (100%). . . 2,337 (54.1%)1,986(45.9%) 1960. . . . . . . . 2,596 (100%). . . 1,142 (44.0%)1,454(56.0%) 1950 to 1959. . . .19,784 (100%). . .10,186 (51.5%)9,598(48.5%) 1940 to 1949. . . .32,317 (100%). . .13,695 (42.4%)18,622(57.6%) 1930 to 1939. . . .14,259 (100%). . . 3,084 (21.6%)11,175(78.4%) 1920 to 1929. . . . . .19 (100%). . . . .17 (89.5%)2(10.5%) Prior to 1920 . . . . .28 (100%). . . . .25 (89.3%)3(10.7%) Unknown . . . . . . 4,780 (100%). . . 3,503 (73.3%)1,277(26.7%) Totals. . . . . . 168,063 (100%). . .59,497 (35.4%)108,566(64.6%) Data source: Bureau of Alcohol, Tobacco and Firearms. Data in the row entitled "Totals," column (2) and all percentages were calculated by Eric M. Larson. Note: The National Firearms Act (NFA) was enacted in 1934, and became effective on July 26 of that year; therefore, all registrations ATF has reported for years prior to 1934 are errors. Of the 14,259 firearms that ATF reports were originally registered during 1930 to 1939, the vast majority (11,449, or 80.3 percent) were Form 1 registrations. In 1968, 54,505 (nearly 93 percent) of the 58,904 registrations were Form 4467 (amnesty period) registrations. The relative lack of NFA firearm transfers is even greater for most Form 4467 registrations. Consider this: of the 58,904 firearms registrations in 1968 (as reported in 1996), nearly all of them (54,505, or nearly 93 percent) were Form 4467 registrations. Of the 58,904 firearms registered in 1968, 50,314 (more than 85 percent) have not been legally transferred since 1968. For all firearms registered during 1934 to 1971, including those registered during unknown years, the data in Table 5 show that 108,566 (nearly 65 percent) are reported by ATF to still be owned as of December 31, 1996, by persons or institutions that lawfully registered or acquired them during those years. 1. Some NFA firearms are apparently currently registered to persons whom ATF has stated are deceased. Some of these 108,566 firearms are registered to persons known to be deceased. Indeed, in 1981, an internal ATF study reported: We have the condition where people who registered firearms under the original National Firearms Act at age 65 would now be 112 years old. We know that these people are dead and their heirs have not taken the necessary steps to contact us so that the involuntary transfer created by the registrant's death can be formalized. Of the 14,259 firearms registered during 1934 to 1939, ATF reports that 11,175 (more than 78 percent) are still registered to the same persons or institutions that acquired them during that period. Consider even the most conservative possible scenario: A person who was 21 in 1939 (the minimum age for lawful ownership of an NFA firearm) would turn 79 sometime in 1997. How many of these people are still alive today? A person who was 65 in 1934 would have been born in 1869, and a person who was born in 1910 would be 87 years old in 1997. Is it safe to assume that nearly all NFA firearms currently registered to persons who were born during 1869 to 1910 alone are no longer actually possessed by these persons? It appears that the problem of NFA firearms registered to persons who are deceased is consistent, particularly for two of the major time periods that private citizens voluntarily registered large numbers of unregistered NFA firearms. Specifically, the proportion of NFA firearms still owned by the same persons or institutions since they were registered by or transferred to them in 1968 is more than 85 percent as of the end of 1996, which is very similar to the pattern observed for 1934 to 1939 (more than 78 percent). A person who was 21 years old in 1968 would turn 50 sometime in 1997, and someone who was 65 would be 94 years old. Again, it is likely that at least some of these persons are deceased. As the owners of these lawfully registered NFA firearms die, physical custody of the firearms may pass to a variety of persons. In some instances, the firearms involved may not obviously be "gangster weapons," such as a hand grenade or a machinegun. Firearms classified as "Any Other Weapon," such as animal trap guns, certain .410 handguns, and various novelty or gadget-type devices (i.e., a knife-pistol) may be regarded as collector's items rather than guns. Statistical Probabilities Demonstrate that on Average, the Attempted Transfer of a Registered NFA Firearm for Which ATF May Have Lost the Registration Would be a Relatively Rare Event Some example calculations demonstrate that the average probability that an attempt to transfer ownership of a registered NFA firearm for which there is no record in the NFRTR would be quite small a relatively rare event. These calculations represent the average probabilities involved and involve several major assumptions: 1. That ATF lost or destroyed 121 original registrations for NFA firearms registered during 1934 to 1939, that the owners didn't lose their copies of these registrations, and that all 121 firearms were detected when the owners attempted to transfer these firearms sometime during 1940 to 1996. Further assume that the number of these "lost then found" firearm registrations is equal for each year. These 121 hypothetical firearms represent about 0.85 of 1 percent (that is, less than 1 percent) of the 14,259 NFA firearms registered during 1934 to 1939. 2. That ATF lost or destroyed 500 original registrations for NFA firearms registered in 1968, that the owners didn't lose their copies of these registrations, and that all 500 firearms were detected when the owners attempted to transfer these firearms sometime during 1969 to 1996. Further assume that the number of these "lost then found" firearm registrations is equal for each year. These 500 hypothetical firearms represent about 0.85 of 1 percent (that is, less than 1 percent) of the 58,904 NFA firearms registered in 1968. These assumptions imply that a very small number of "lost and found" NFA firearms would show up every year, even given that I have assumed that ATF has lost or destroyed as many as 621 registrations. Consider the 500 hypothetical firearms registered in 1968, for which I assume ATF lost or destroyed the registration documents. The average number of firearms that would turn up each year from 1969 to 1996 (28 years) is based on 1 in 28 (1 28 = .0357143), or nearly 3.6 percent per year for each of the 28 years between 1969 and 1996, and is still a relatively small number. Specifically, 500 firearms x .0357142 percent = 17.8571, or about 18 firearms per year. Note that 3.571 percent x 28 years = 100 percent. Putting it another way, 17.8571 firearms x 28 years = 500 firearms. For the 121 hypothetical firearms registered during 1934 to 1939, their equal distribution every year is based on 1 in 58, corresponding to the number of years available for transfer from 1940 to 1996. The yearly average that one of these 121 hypothetical firearms would show up during an attempted transfer would thus be 1 in 58 (1 58 = .0172413), or slightly more than 2 firearms per year (121 firearms x .0172413 percent = 2.0861973, or about 2.1, firearms). Again, note that 2.0861973 firearms x 58 years = 121 firearms. These calculations imply that transfers are attempted for an average of 20 registered firearms per year (17.9 + 2.1 = 20) for which ATF has no record, for the period from 1969 to 1996. These are limited to firearms registered during 1934 to 1939, and during 1968, on Form 1 or Form 4467, as previously explained. While these calculations identify an estimated average number of registered firearms for which ATF has no record which are sought to be transferred each year, I would caution anyone from considering these probabilistic calculations to be anything other than a hypothetical illustration of the process. The reason is that some of the assumptions are highly unlikely for one thing, I possess no conclusive evidence that ATF has lost or destroyed 621 registrations of these types. I have also assumed that each firearm for which ATF lost or destroyed the original registration has an equal probability of being transferred each year, and that is most certainly not true. Some firearms are owned by federal, state or local government agencies, including museums, and are never sold. Others remain in the same families for many years and are transferred, if at all, only upon the death of the original registrant. Deaths to original registrants of NFA firearms would also logically be expected to vary according to how old the person was at the time he or she registered the firearm; consequently, the timing of what ATF terms these "involuntary" transfers would not be uniform over time, to the extent that these firearms are registered to individuals. It would be extremely difficult and perhaps impossible to calculate the true probabilities, because separate calculations would have to be performed for each year using the exact number of firearms transferred, and adjusted for possible additions of registered firearms for which ATF had no record to the overall data base. It is likely that the data necessary to actually perform all these calculations for all years probably do not exist. What I have calculated are yearly averages, because I do not have access to other data at this time. Consider, however, the results of my hypothetical estimates with the actual number of firearms ATF has apparently added to the NFRTR each year during 1992 to 1996, as shown in Table 6. The data sets I used to calculate each set of estimates are not identical, but each are based on actual NFRTR data. My hypothetical estimates are based partially on NFA data and, as earlier discussed, rely upon some extremely unlikely assumptions. Table 6 about here Table 6 Numbers of Firearms that ATF Has Apparently Added to the National Firearms Registration and Transfer Record during 1992 to 1996, versus Hypothetical Estimates For These Same Years Calculated Using Actual NFRTR Data for Each of the Years Shown Registration 1993 1994 1995 1996 TotalsAnnual averages Hypothetical (5) 4 years = estimates (1) (2) (3) (4) (5) (6)(7) Form 1. . . . 2 7 20 7 . . .369.00 per year2.1 per year Form 4467 . .-1 23 43 18 . . .8320.75 per year17.9 per year Totals . . .1 30 63 25119 29.75 per year20.0 per year Data source: The data in columns (1) through (6) are taken from Table 3 of this article, and were calculated using actual NFRTR registration data for 1992 to 1996, as reported by ATF. The year 1992 is not listed here because it is the base year from which these comparisons originate. The data in column (7) are hypothetical estimates calculated by Eric M. Larson, under the assumptions that: (1) ATF lost 121 or destroyed Form 1 registrations during 1934 to 1939, 500 Form 4467 registrations in 1968; (2) the transfers of firearms corresponding to the lost or destroyed registrations were equally distributed across the years since their date or period of original registration (since 1940 and 1969, respectively); (3) the original owners of these firearms all kept their registration documents; and (4) when confronted with these original documents, ATF was forced to add the firearms to the NFRTR. The data I used to estimate that the 20 firearms added per year as reported in column (7) of Table 6 are based on (1) two data sets: (a) the total number of firearms registered during 1934 to 1939, and (b) the total number of firearms registered during 1968; and (2) the assumption that ATF lost 121 Form 1 registrations from 1934 to 1939, and 500 Form 4467 registrations from 1968. These data sets are different (but not all that different) from the actual NFRTR data that I used to calculate that 30 firearms were added per year, as shown in column (6). The data in column (6) are based on (A) all Form 1 registrations from 1934 to 1971 (and unknown years) as reported by ATF during 1992 to 1996 (see Table 2), an average of 32,728 firearms; and (B) all Form 4467 registrations, including those for "unknown" years, averaging 57,202 (see Table 1). The average of 20 guns per year that I hypothetically estimated would be added to the NFRTR in the Form 1 and Form 4467 categories together is just that a hypothetical average annual estimate. The actual data would be expected to change from year to year, as they do in columns (1) to (4) of Table 6, and the true numbers may be larger than shown because of underestimations. My hypothetical estimates (i.e., 20 per year) clearly do not constitute legal evidence that ATF has lost or destroyed 621 NFA firearm registration forms, but that wasn't their objective. My intention was to create a simple model to demonstrate how some number of registrations that ATF may have lost or destroyed could show up as "new" registrations as owners who kept copies of their registrations seek to transfer ownership of their firearms. This does, in fact, appear to be what has actually happened. The reason is that the actual NFRTR data demonstrate that about 30 NFA firearms per year, registered on Form 1 or Form 4467, were apparently added to the NFRTR data base during 1992 to 1996. This series of estimates for 1992 to 1996 isn't long enough to test whether the annual average of nearly 30 is the true average for the period from 1969 to 1996, to which it applies. Also, the number of Form 1 and Form 4467 registrations that were added during 1992 to 1996 are probably underestimated. The data in columns (1) to (4) of Table 6 do, however, exhibit the kind of natural variation that would be expected to occur as owners of lawfully registered NFA firearms die at different rates and times. Total registration activity for all years was 2,340,578 transactions as of December 31, 1996, and 621 is a very small fraction that is, .02653 of 1 percent of 2,340,578. In relative terms the number is small, but 621 violations of the NFA add up to a lot of potential criminal prosecutions. Each conviction for a single nonregistration violation can mean a fine of up to $250,000 and up to 10 years imprisonment. The Integrity of the NFRTR Data are Questionable Because it Appears that ATF May Have Registered Some NFA Firearms in Violation of the NFA Many people who collect or deal in NFA firearms have heard rumors, or believed, that some ATF agents have registered unregistered NFA firearms, i.e., "You can get [a contraband firearm] registered if you know somebody." I've heard comments like these for years from attorneys, firearms dealers, and collectors who are "positive" that it's happened. Beliefs, no matter how strongly felt, however, do not constitute valid and reliable evidence. Recently, Mr. Sanders told me he believed that the illegal registration of an NFA firearm by an ATF agent was a virtually impossible event. "Absolutely not," he told me. "I just don't believe that any ATF agent would risk that, because it would be discovered and there would be a trail right back to him." Inspecting some transfer documents and other evidence has caused me to be quite suspicious of the circumstances of some registrations. For me, a definitive conclusion in this area is not possible at this time. Consider the following three situations. 1. Do any of the 146 Form 4467 registrations that occurred during 1972 to 1986, as reported by ATF in 1996, correspond to firearms that ATF has registered illegally? As we have seen, NFA Branch firearms specialist Gary N. Schaible told me that ATF accepted and approved Form 4467 (amnesty period) registrations from U.S. military personnel who returned to the United States following an overseas tour of duty until 1971. An example of such a registration, which was submitted on March 3, 1969, and accepted for registration on March 14, 1969, has been inserted following this page. Insert page about here The physical approval of this Form 4467 was accomplished through the use of a facsimile or rubber stamp device bearing the typed name of Harold A. Serr, who was then ATF Director. The notation "Signed---Harold A. Serr" apparently was legally sufficient to serve as evidence of the legal registration of this particular firearm. Mr. Schaible has stated that ATF accepted applications to register unregistered NFA firearms after the amnesty period expired on December 1, 1968. It is important to identify the legal authority under which these registrations were accepted, because an "IMPORTANT" statement on Form 4467 itself states that the ATF Director "cannot" accept Form 4467 to register a firearm "except when received by Director during the time period November 2, 1968, through December 1, 1968." Did the Congress authorize these post-December 1, 1968, registrations? If so, when? Were regulations by ATF or some other entity describing how Form 4467 registrations could be made published in the Federal Register? Can anyone at ATF identify any provision of law which would have allowed ATF to accept these registrations? I was intrigued by the post-1971 Form 4467 registrations, and asked Mr. Schaible several questions about them when I interviewed him in April 1996. These registrations looked mostly like errors, he said, and agreed that it would have been illegal for ATF to register NFA firearms after 1971. "What about the 36 that are shown for 1976?" I asked. "What about those? And the 85 registered in 1972?" "I'm not sure," he said." "You mean, you can't explain them?" I asked. "That's what I'm saying," he answered. "I have no explanation for them." There have been many people who missed the opportunity to register unregistered NFA firearms during the original amnesty period, which lasted from November 2, 1968, to December 1, 1968, in part because ATF failed to explain that certain firearms (such as smooth bore pistols, and animal trap guns) were required to be registered. Did ATF register unregistered NFA firearms on Form 4467 after April 5, 1971? The registration of unregistered NFA firearms by ATF after the amnesty period expired raises troubling questions about the custom and culture of ATF, and whether such practices have extended to the future. I believe that all of the Form 4467 registrations that ATF accepted after December 1, 1968, were clearly in violation of Congressional intent and probably in legal violation of the NFA. Also consider the statements regarding an amnesty period by ATF agent Cecil M. Wolfe to executives of the Thompson-Center Arms Co. in Rochester, New Hampshire, on June 18, 1969, regarding the status of rifled-barreled .45/.410 Thompson Contender pistols. Mr. Wolfe threatened to declare these single-shot handguns subject to the NFA, although under the NFA (and according to a February 11, 1969, Memorandum from ATF Director Harold A. Serr), they were not subject to the NFA. Mr. Wolfe nevertheless stated: Since the amnesty period is over there could not be registration. They would be contraband and subject to seizure. The owners would be in violation. It seems to me that Mr. Wolfe's statement, juxtaposed with the fact that ATF accepted registrations of unregistered NFA firearms after the amnesty period expired (according to Mr. Schaible, and the Form 4467 registration that was accepted on March 14, 1969, that I have reproduced), is a double standard of dubious (il)legality. But for now, let's examine the Form 4467 registration data that ATF has reported during 1992 to 1996, against the legality of these registrations by year of registration. I have reworked Table 1 (see page 6 of this study) to emphasize the out-of-date range Form 4467 registrations, as shown in Table 7 all the data are the same. All of the registrations that are logically impossible (pre- 1968) or definitely illegal (post-1971) have been emphasized in bold-face type. All of the registrations that are mostly legal (1968 to 1971) have been italicized. I have left registrations for "unknown" years, and the rows labeled "Totals" and "Change from previous year," in regular type. Table 7 about here Table 7 Form 4467 (Amnesty Period) Registrations as Reported by ATF in the National Firearms Registration and Transfer Record During 1992 to 1996: Pre-1968 (None are Possible), 1968-71 (Most are Legal) and Post-1971 (All are Illegal) Net change Net number Year of in Form 4467 of firearms original registrations, added or Form 4467 Years of registration activity 1992-1996 removed, registration 1992 1993 1994 1995 1996 (6) - (2) = 1992-1996 (1) (2) (3) (4) (5) (6)(7) (8) 1988 1 0 1987. . . . . 1 1 1 1 -1 -1 1986. . . . . 3 3 3 3 1 -2 -2 1985. . . . . 1 1 1 1 -1 -1 1984. . . . . 3 3 3 3 -3 -3 1983. . . . . 1 1 1 1 -1 -1 1982. . . . . 3 3 3 3 -3 -3 1981. . . . . 1 1 1 1 -1 -1 1980. . . . . 4 4 4 4 -4 -4 1979. . . . . 1 1 1 1 -1 -1 1978. . . . . 6 6 6 6 -6 -6 1977. . . . . 1 1 1 1 1 0 1976. . . . .39 39 39 39 36 -3 -3 1975 1974. . . . . 5 5 5 5 5 0 1973. . . . .17 17 18 18 18 1 +1 1972. . . . .84 84 84 84 85 1 +1 1971. . . . .26 26 26 26 26 0 1970. . . . 272 272 271 271 271 -1 -1 1969. . . 2,006 2,010 2,016 2,017 2,053 47 +47 1968. . .54,48754,485 54,485 54,503 54,505 18 +18 1967. . . . .64 64 64 64 64 0 1966. . . . . 8 8 8 8 8 0 1965. . . . . 2 2 2 2 2 0 1964. . . . . 3 3 3 3 3 0 1963. . . . . 2 2 2 2 2 0 1962. . . . .14 14 14 14 14 0 1961. . . . . 2 2 2 2 2 0 1960. . . . . 6 6 6 6 6 0 1950-59 . . .23 23 23 23 23 0 1940-49 . . . 9 9 9 9 10 1 +1 1930-39 . . .25 25 25 25 27 2 +2 1920-29 . . . 6 6 6 6 3 -3 -3 before 1920 . 4 4 4 4 -4 -4 Unknown . . .58 58 57 57 58 0 Totals. .57,18757,189 57,196 57,216 57,223 36 +36 Data Source: Bureau of Alcohol, Tobacco and Firearms. NOTE: Registration data shown in boldface represent registrations that are not possible (pre-1968) or illegal (post-1971). Form 4467 did not exist until 1968; Form 4467 registrations were prohibited by the U.S. Supreme Court on April 5, 1971 (United States vs. Freed, 401 U.S. 601, (1971)). Registration data shown in regular type represent Form 4467 registrations that are probably mostly legal. Blank spaces represent zero. Why would somebody at ATF reclassify 26 post-1971 Form 4467 registrations in 1996, but virtually none of the pre- 1968 registrations? I don't know. I wouldn't presume to guess. But if somebody at ATF spotted a post-1971 Form 4467 registration, it would be illegal for them to change it. Intriguingly, the total number of Form 4467 registrations actually increased from 57,216 to 57,223 during 1995 to 1996. That's an increase of seven more guns. Where did these firearms come from? The number of post-1971 registrations ATF reported in 1995 decreased from 172 (26 registrations) to 146 in 1996, which may indicate reclassification in response to my criticisms. If so, why not correct (reclassify) the other post-1971 registrations? In my judgement, Ms. Levine and other officials of the NFA Branch have been more than aware of my comments about post-1971 Form 4467 registrations listed in the NFRTR data. How could they not be, given that I made my comments part of the official record of ATF's appropriations hearing? So, why not correct the post-1971 Form 4467 data errors, if they are indeed errors? Answer: Because they aren't errors; these data actually represent firearms that somebody at ATF registered illegally and, interestingly, the NFA prohibits any change in a record other than to correct an error. Attempting to cover up an illegal NFA registration by claiming it was a data error would be a serious crime. Heavy legal penalties would apply to any person, including any ATF agent or employee, who knowingly made or caused the making of a false entry on any NFA record (see Title 26 U.S.C.,  5861(l) and  5871). 2. Did ATF illegally register a 28 gauge H&R Handy-Gun bearing serial number 29691 in 1972? A number of years ago I met Bob Miller of Houston, Texas, a long-time collector who owned more than 100 NFA firearms. Among the firearms he owned was a 28 gauge H&R Handy-Gun, which is quite a prize. In fact, a different collector I know ran advertisements in every issue of The Shotgun News for about six years during the 1980s, offering to buy H&R Handy-Guns. During that entire time, he said, nobody ever offered him even one in 28 gauge. In 1985, he was finally able to purchase one from a retired district attorney for $650, and he considered the price a bargain. "I paid that $650 gladly, by the way," he told me in 1987, when I gasped at the cost. "I don't know where I'd ever find another one," he added. "I don't know that I could ever replace it. It really doesn't matter how much money you have the guns are simply not out there, not the 28 gauge ones." Bob Miller who said he had been an NFA firearms manufacturer for some 30 years told me that he didn't register the 28 gauge H&R Handy-Gun he owned during the 1968 amnesty. The reason, he said, is that he was afraid that the amnesty was a pretext to confiscate guns. A couple of years later he said he regretted the decision, and had heard through "the grapevine" that a returning serviceman could register such firearms on Form 4467 and that ATF would approve them. "So," he said, "I asked around and found a guy to register it for me, and to transfer it right back to me." So he came to legally own the 28 gauge. Mr. Miller owned several other legally registered H&R Handy-Guns, and had no intention of selling any of them. Mr. Miller also probably never thought that he was going to die in 1993 and that I would wind up buying several H&R Handy-Guns from his estate. Something like 6 or 8 months after taking lawful possession of the firearms, I got a surprise in the mail: The person who settled the estate decided "to just send you every piece of paper I had with 'H&R Handy-Gun' on it, because of your interest as a collector," as I found out when I called to give thanks. These papers included a brand new 1932 Harrington & Richardson Arms Co. catalog (which listed the H&R Handy- Gun) in its original envelope, and some very interesting paperwork. As we have seen, the U.S. Supreme Court prohibited ATF from allowing private citizens to voluntarily register unregistered NFA firearms on April 5, 1971. ATF recently confirmed that this particular 28 gauge H&R Handy-Gun, bearing serial number 29691, was originally registered on March 2, 1972 more than a year after the Court's decision. The interesting question here, obviously, is whether ATF registered this firearm illegally, which is prohibited under Title 26, U.S.C.,  5861(e) and/or (l), and  5871, in light of the 1971 decision. The paperwork I obtained includes the original Form 4 that was used to transfer the 28 gauge H&R Handy-Gun to Bob Miller, and is interesting for several reasons. Insert Form 4 about here 1. Unlike all the other approved Forms 4 that I possess for NFA firearms I lawfully obtained by transfer, this particular approved Form 4 has an original signature in blue ball-point pen ink. In contrast, ATF has placed a stamped facsimile signature on all of the Forms 4 that I signed, which are the only evidence that I possess indicating that I am the legal owner of these firearms. The significance of an original signature, in my judgement, is that nobody at ATF could ever claim that "somebody" illegally registered an NFA firearm through the unauthorized use of a facsimile signature device. 2. The person who signed this particular Form 4 wasn't a statistical clerk. It was Rex D. Davis. On March 23, 1972, when he approved this Form 4, Mr. Davis was the Director of the Bureau of Alcohol, Tobacco and Firearms. I'm not aware of a higher authority at ATF than its Director. 3. This particular Form 4 could be forensically examined. The signature of the transferor, Alan C. Buck, could be compared with the original NFA firearm registration document that ATF (theoretically) has in its files to back up the information in the NFRTR, as well the signature of the person who accepted it for registration. Mr. Davis' signature on this Form 4 appears to be genuine. Finally, two additional elements of fact make this case even more interesting. One fact is a provision that "any firearm involved in any violation of the provisions" of the NFA "shall be subject to seizure and forfeiture" under Title 26, U.S.C.,  5872(a). The other fact is that: Purchasers of registered National Firearms Act firearms . . . have no legal or practical means of determining the pedigree of a registered firearm and are totally at the mercy of BATF's approval of the transfer application and registration (BATF Form 3, 4 or 5) by which the purchasers obtain authority to receive and possess the firearm. BATF refuses to disclose to subsequent registrants the prior registration and transfer forms pertaining to any National Firearms Act firearm, citing the taxpayer privacy provisions of the Internal Revenue Code, 26 U.S.C. section 6103. Thus, purchasers/transferees of National Firearms Act firearms are totally at the mercy of BATF's competence and diligence, or lack thereof, in obtaining valid and permanent possession of a validly registered firearm, and in being able to subsequently effect a legal transfer of such firearm. By their very nature, legally restricted and often of historical significance, National Firearms Act firearms ordinarily are valued at thousands of dollars each. There is potential for conflict here. It is reasonable to assume that because ATF approved the transfer of the 28 gauge H&R Handy-Gun bearing serial number 29691 to me on August 22, 1994, on a Form 4, I lawfully possess this firearm. What if, more than 25 years ago, the Director of ATF may have personally approved the transfer of an NFA firearm that was illegally registered? This example is very unusual because, as previously noted, a person who legally buys a registered NFA firearm as I did normally has no way to find out anything about the registration history of that firearm. 3. Why were five firearms originally manufactured in or before 1934 not registered by the manufacturer until April 16, 1986? Consider the example of NFA firearms being originally registered in 1986 by a manufacturer 52 or more years after they were originally manufactured. The firearms in question are H&R Handy-Guns bearing serial numbers 5592, 43950, 50885, 52551, and 53637. I know something about these firearms, because I lawfully own three of them. On January 13, 1997, ATF confirmed that they were originally registered on April 16, 1986. The other two (serial numbers 43950 and 52551) also were originally registered on April 16, 1986. I don't know all the technical "ins and outs" of what a manufacturer must do to register an NFA firearm. My impression is that such firearms have to be registered soon after they are actually "made." What's interesting about these five guns is the fact that the Harrington & Richardson Arms Co. advised ATF in writing on November 27, 1953, that "H & R has not manufactured Handy Guns since the {NFA] law was passed in 1934, classifying these guns in the same category as machine guns." The letter goes on to state that "in the last two (2) years, all our Handy Guns in .410 gauge and 28 gauge were exported to Canada." According to several long-time H&R employees, four of these five H&R Handy-Guns were kept in a small factory collection at Harrington & Richardson Arms Co. since their original manufacture. The fifth serial number 53637 was added during the 1970s, when H&R purchased it to add to the factory collection. Another, serial number 5592, is a unique factory prototype in pristine condition and in its original box; and serial number 50885 is new in its serial-numbered original box, and apparently has never been fired. The blueing on the butt end of the barrel is completely intact. It is difficult for me to imagine that Harrington & Richardson Arms Co. didn't register these guns with ATF in 1934. The NFA required their registration, and to a firearms manufacturer the price of noncompliance would have brought heavy penalties. If they weren't registered in 1934, they certainly must have been registered in 1968 again, as firearms manufacturers, the executives of Harrington & Richardson were certainly aware of the new provisions under the Gun Control Act of 1968. It is difficult for me to believe that these executives would not have made sure that the company totally complied with the law. In my opinion, Harrington & Richardson Arms Co. probably did comply with the law in 1934 and that ATF and H&R each lost or destroyed the original registration documents. I certainly cannot prove that. But I can prove that H&R stopped manufacturing the H&R Handy-Gun in 1934, and that all five guns noted above were originally registered by ATF on April 16, 1986. Some Current (1992 to 1996) NFRTR Data are of Questionably Accuracy and Integrity, and the Meaning of Some of These Data is Unclear To try and understand how the NFA database is constructed, I filed a Freedom of Information Act (FOIA) request on November 21, 1996, to obtain the ATF manuals describing how NFA information is entered into the NFRTR. My request stated that I was not requesting any privileged taxpayer information, only "ordinary coding and data entry information that would be necessary for a computer operator to enter and/or retrieve data from the NFRTR." On January 13, 1997, ATF denied my request in total, and stated that the information I requested is exempt from disclosure because it is "related solely to the internal personnel rules and practices of an agency." I discuss this FOIA later in this study. It is hazardous to make final, definitive judgements about the validity and reliability of data in the NFRTR or, indeed, any computerized data base, without (1) inspecting the data entry manuals to determine how data are defined, entered, and edited; (2) interviewing personnel who are responsible for entering and retrieving these data, as well as all other personnel who have access to the NFRTR; and (3) independently forensically auditing, as necessary, samples of NFRTR data and/or specific records, including matching computerized data entries to paper documents. An independent forensic audit would include: (1) tests for the validity of all current registrations or transfers; (2) a determination of the extent, if any, to which firearms are currently registered to persons who are deceased; and (3) as required, the scientific examination of questioned documents. Editing, correcting, "cleaning" or "purifying" a data base are common, standard and necessary procedures, because data entry errors, and mechanical "system" errors, can and do occur routinely. In themselves, data entry (sometimes termed "keypunch") errors are not sinister, nor are system malfunctions, or software bugs. Because of the serious legal penalties associated with violations of the NFA, however, a data base such as the NFRTR should be virtually free of errors, including misclassifications. To try and get a better idea of the magnitude of the possible errors in the NFRTR for all data from 1934 to 1971, I inspected total "REGISTRATION ACTIVITY" for these years (and unknown years) reported by ATF from 1992 to 1996. These data are shown in Table 8. Table 8 about here Keep in mind that virtually all of the data in Table 8 are a report by ATF of registration activity that is supposed to have largely occurred from 25 to 62 years ago, and that the changes in these data occurred sometime during 1992 to 1996. As Table 8 shows, the "REGISTRATION ACTIVITY" for most years has been fairly stable, and it is possible that changes in numbers of registrations were made for the reasons NFA Branch Chief Nereida W. Levine stated in her January 9, 1997, letter. Ms. Levine stated that there are "errors in the date or form fields," and "errors, blanks, or misrepresented characters in the date field." Ms. Levine further wrote, as we have seen: If we do locate a record where the date, form number, or other information was not entered correctly, we enter the correct information. These actions may then result in an adjustment to previously generated statistics. We would like to point out that errors in the date or form number fields do not affect the thoroughness of a search of the database by NFA Branch personnel. Consider Ms. Levine's statements in light of the LeaSure case, where NFRTR custodian Gary N. Schaible testified that ATF's records did not show that certain firearms were registered to Mr. LeaSure and, therefore Mr. LeaSure was in illegal possession of the guns. As we have seen, U.S. District Judge John A. MacKenzie threw out all of Mr. LeaSure's convictions for nonregistration of firearms. An attorney familiar with the case later wrote: In essence, Schaible was testifying that "We can't find an official record and therefore the defendant is guilty." What we know now is that Schaible should have testified that "We can't find half our records even when we known they're there and therefore we're not sure if anyone is guilty." Table 8 Total "REGISTRATION ACTIVITY" from 1934 to 1971, and During Unknown Years, as Reported by ATF in the National Firearms Registration and Transfer Record During 1992 to 1996 Net number of registrations added or Years of registration activity removed, Year 1992 1993 1994 1995 1996 1992 to 1996 (5) - (1) = (1) (2) (3) (4) (5) (6) 1971. . . .5,086 5,087 5,086 5,092 5,0959 1970. . . .2,168 2,168 2,168 2,168 2,167-1 1969. . . .4,065 4,071 4,077 4,079 4,11449 1968. . . 58,844 58,845 58,844 58,877 58,90460 1967. . . .3,465 3,464 3,461 3,461 3,461-4 1966. . . .3,856 3,856 3,856 3,857 3,8582 1965. . . .3,736 3,737 3,737 3,739 3,7393 1964. . . .2,808 2,808 2,805 2,805 2,804-4 1963. . . .2,672 2,672 2,672 2,672 2,6720 1962. . . .3,135 3,136 3,143 3,143 3,1438 1961. . . .4,321 4,322 4,319 4,320 4,3232 1960. . . .2,592 2,594 2,595 2,596 2,5964 1950-59 . 19,771 19,769 19,766 19,778 19,78413 1940-49 . 32,307 32,309 32,301 32,307 32,31710 1930-39 . 14,250 14,250 14,248 14,252 14,2599 1920-29 . . . 53 54 54 54 19-34 before 1920 .121 121 122 120 28-93 Unknown . .4,147 4,493 3,944 4,274 4,780633 Totals. .167,397 167,756 167,198 167,594 168,063666 Change from previous yearn/a +359 -558 396 469666 Data Source: Bureau of Alcohol, Tobacco and Firearms. All numbers shown in boldface type were calculated by Eric M. Larson. Note: The National Firearms Act (NFA) was enacted in 1934, and became effective on July 26 of that year. All registrations ATF has reported for years before 1934 are errors. After inspecting them, I decided that the data in Table 8 raise more questions than they potentially resolve, for several reasons: 1. The data ATF reports do not appear to reflect changes in total registration that could be expected to occur involving NFA firearms which ATF has removed from the NFA as collector's items. As with Form 1 and Form 4467 registration data, it could make sense for the total number of registrations to decrease as a consequence of ATF removing certain firearms from the NFA as collector's items. In a letter dated March 3, 1997, Joseph J. Vince, Jr., Chief of ATF's Firearms Enforcement Division, stated: "Requests for removal of weapons from the purview of the NFA are frequently granted." Yet, as we have seen, the net total number of registrations has consistently increased each year from 1992 to 1996. Does ATF report or record removals of firearms from the NFRTR? If so, where? When? And how many? As of 1996, there were 57,223 amnesty (Form 4467) registrations, up from 57,187 in 1992; and Form 1 registrations for years in and before 1971, and for unknown years, increased by 83 during this same period. The data in Table 8 indicate that total registrations increased by 666 firearms from 1992 to 1996; and that 633 (95 percent) occurred during "unknown" years. Also consider that 109 firearms apparently were added in the two-year period during 1968 to 1969 when more than 60,000 NFA firearms were registered. Is it not possible that some of those registrations from 1968-69 were lost or deliberately destroyed by ATF? And isn't it further possible that some of the people who registered firearms during 1968-69 didn't lose their original registrations? The data I have analyzed in this article, juxtaposed with the Busey matter and the LeaSure case, surely raises valid doubts about the accuracy and integrity of the NFRTR. In other research, I and other firearms experts have estimated that ATF has already removed 50,000 to 100,000 or more NFA firearms from the NFA as collector's items. To what, if any, extent are these removals reflected in current NFRTR data? At least some of the NFA firearms that ATF has removed from the NFA as collector's items were registered, although some were definitely not registered at the time of their removal. The NFRTR data base instruction manuals may identify some procedure for removing firearms from the NFA as collector's items, and thus shed some light on whether these registrations were removed from the NFRTR data base, as would knowledge of how many individual firearms ATF removed during 1992 to 1996. 2. The significance of increases in the number of registrations for "unknown" years during 1992 to 1996 is unclear, and merits investigation. Consider that the total number of registrations for "unknown" years increased by 359 from 1992 to 1993, decreased by 558 the following year, and increased by 865 during the past two years. There's a logical disconnect here: How can firearms registrations be listed for an "unknown" year, if they were reported as having occurred during that year? This makes no sense. 3. Increases in the overall number of firearms may indicate that ATF personnel have lost or deliberately destroyed original registration documents, and that registrations have been added to the NFRTR whenever a lawful owner was able to produce his or her copy of the registration document. In inspecting the available documentation on the NFRTR, I am struck by the evasive quality of ATF's statements regarding its accuracy. Consider Packet Exhibit #2 in the May 20, 1996, Department of Justice letter. It summarizes an interview with Thomas Busey regarding his comments about the accuracy of the NFRTR. "The scope of the interview was limited" in accordance with a discussion with ATF's Chief Counsel, Mr. Busey's interviewer reports, but does not identify what was "limited." Consider his report of Mr. Busey's explanation: When [Mr. Busey] said that members of this staff testify that the NFRTR database is 100% accurate although they know otherwise, he made a misstatement of the facts. What he meant to convey was the fact that the database contains certain inaccuracies which can be attributed to human error. His personnel testify only to the accuracy and diligence of their search and make no comment, either in court or on any official document, concerning the accuracy of the database. If he were asked about the accuracy of the database, he would reply that the database contains evidence of human error. He would then explain how a search is performed. This admission and, indeed, the content of the "interview," is startling: the accuracy of the NFRTR data base is unknown, and ATF personnel testify only about how hard they looked in a data base that contains inaccuracies. Perhaps the most blatant and incredible omission is that no one has apparently asked Mr. Busey, under oath, when and from whom Mr. Busey said he "heard" (i.e., "I've heard that's happened") that registrations for firearms ATF agents thought were unregistered were waved in the agents' faces by irate lawful owners. Mr. Busey's statement is even more jarring when viewed against statements in an April 30, 1991, Memorandum by Wayne Miller, then Chief of the NFA Branch, who declared: "As part of the implementation of the automated [data] retrieval system, we have verified the accuracy of the information in the NFRTR." It appears that, in 1996, ATF has decided to retreat from its 1991 statement, confirm that "the database contains evidence of human error" and deliberately not be specific as to exactly what "human error" means. One can hardly be comforted by Mr. Miller's quotation of a 1980 court decision regarding the accuracy of the NFRTR: Considerable evidence was received that [ATF's] officials have for many years recognized the inadequacy and incompleteness of the Bureau's records. The Court is not required to pass judgement on this, however, because the Government has failed to show that these seven items are firearms. Mr. Miller also stated that while "the inadequacy and incompleteness" of the NFRTR data reflected conditions that apparently were true when the criticisms were made, these conditions "had largely been corrected" by 1980. 4. My direct questions to NFA Branch personnel asking whether ATF has added or would add firearms to the NFRTR after a lawful owner produced a registration or transfer document of which ATF had no record, did not result in any ringing denials. "Has ATF ever added a firearm to the NFRTR, after a lawful owner produced a valid registration, because ATF had no record of the firearm in the NFRTR?" I asked Mr. Schaible in April 1996. "I assume that's happened," he told me. I asked him again, to make sure I had correctly understood his answer. His response was identical. Ms. Levine, noting that I "asked whether a firearm would be added to the Registry if a person possessed a valid registration that is not in the Registry," stated: "[The firearm] would be added to the National Firearms Registration and Transfer Record if the information was not already in the Record." I had expected a rejoinder (i.e., "but that could never happen, because our records are accurate beyond a reasonable legal doubt"), and there was none. My, my. Reading Ms. Levine's letter was a telling moment: The dog that does not bark. It would seem advisable to obtain sworn statements, under penalty of perjury, from Mr. Busey, Mr. Schaible, Ms. Levine, and all other NFA Branch personnel regarding their personal knowledge or experience regarding the registration or re-registration of NFA firearms, and of lost or destroyed registration, transfer or other NFA documents. 5. The basis of certain NFA firearms "REGISTRATION ACTIVITY," particularly in the "OTHER" category, is unclear or unknown. The ATF data table entitled "NFA REGISTRATION ACTIVITY - ANNUAL COMPARISON" contains two columns of data entitled "LTR" and "OTHER" whose meaning is not well understood. Again, it is possible that access to the NFRTR data entry manuals could help resolve this question. Mr. Schaible told me that "LTR" means a "Letter Registration." A letter registration is when ATF registers a firearm on the basis of a letter that somebody writes to them and, Mr. Schaible said, has been used in cases when ATF forms were unavailable. Are letter registrations just letters from police departments asking that firearms be registered, or do they include testimonials from persons who state they have lost their registration form and would like to re- register their firearms or who simply wished to register unregistered firearms at ATF's discretion? The "OTHER" data are even more intriguing. Mr. Schaible told me that these data resulted from somebody entering the wrong form number into the computer, and the computer automatically routing it to that category. In her January 9, 1997, letter to me, Ms. Levine stated that the "OTHER" category is "comprised of registrations where the form number is different from the other ones tabulated. An incorrect form number would be counted in that column." I wrote to Ms. Levine on January 16, 1997, asking her if she would be kind enough to provide me with a list of the form numbers and the names of the forms, because I was unaware of any other forms that could be used to register NFA firearms; however, Ms. Levine declined to do so. Why? Are the names and numbers of these forms secret? If so, under what authority? What is going on here? Table 9 about here As reported by ATF, "OTHER" data appear in the NFRTR statistics for every year through 1994, including years before 1934 and for "unknown" years, as shown in Table 9, for data reported from 1992 to 1996. What is the actual meaning of these data? Do they represent original registrations, re-registrations, or transfers of registered (or of re-registered) firearms? Consider the fact that the 1980 ATF memorandum I cited which discusses "docket numbers" also describes ATF forms that were used for registrations. Specifically, the 1980 memorandum states that: "the forms are for the returns [registrations] used during the years 1934-1945" and consist of Form 1, Form 2, Form 3 Form 4, Form 5 and Form 6. No other forms are mentioned or identified. Yet, look at what the 1996 data indicate: For example, there were 1,247 "OTHER" registrations during "1930 to 1939"; and during 1980 to 1989, there were 3,127. Indeed, as of 1996, there were a total of 8,876 "OTHER" registrations a far from inconsequential number. I have been researching the NFA for some 10 years, and I've never heard of any forms numbered other than 1, 2, 3, 4, 5, 6, 9, 10 or 4467 being used to register or transfer NFA firearms. According to Ms. Levine, other forms and form numbers exist, but she is apparently unable or unwilling to identify them. What kind of "registration activity" has taken place from 1934 to 1994? What is going on here? 6. ATF personnel identified six types of "Significant Errors" in the NFRTR by June 19, 1995, and inexplicably downgraded them to "Errors" by October 10, 1995, although all are clearly major errors that affect the accuracy and integrity of the NFRTR. Now, recall what Mr. Miller stated in his 1991 memorandum: "As part of the implementation of the automated retrieval system, we have verified the accuracy of the information in the NFRTR." Mr. Miller's choice of exculpatory language in the 1991 memorandum seems particularly insouciant in view of the recent NFA Branch "March 1995 project to clean-up (review and amend) the entire National Registry." Table 9 Number of "OTHER" Category "REGISTRATION ACTIVITY" from 1934 to 1996, and During Unknown Years, as Reported by ATF in the National Firearms Registration and Transfer Record During 1992 to 1996 Net number Year of Net change of firearms original in "OTHER" added or "OTHER" Years of registration activity registrations, removed, registration 1992 1993 1994 1995 1996 1992- 1996 1992-1996 (6) - (2) = (1) (2) (3) (4) (5) (6) (7) (8) 1996 1995 1994 1 1993 25 19 18 14 14+14 1992. . . . 30 25 21 21 19 -11-11 1991. . . . 35 33 29 25 22 -13-13 1990. . . .134 132 130 130 129 -5 -5 1980-89 .3,190 3,179 3,150 3,135 3,127 -63-63 1970-79 . .230 226 222 219 212 -18-18 1960-69 . .131 130 127 125 125 -6 -6 1950-59 . . 47 47 46 45 45 -2 -2 1940-49 . . 57 57 56 55 54 -3 -3 1930-39 .1,280 1,280 1,268 1,263 1,247 -33-33 1920-29 . . .9 10 9 9 5 -4 -4 before 1920 15 15 15 13 5 -10-10 Unknown .3,513 3,749 3,159 3,377 3,872 359+359 Totals. .8,671 8,908 8,252 8,435 8,876 205+205 Change fromn/a +237 -656 +183 +441 205+205 previous year Data Source: Bureau of Alcohol, Tobacco and Firearms. The data in the rows entitled "1960-69," "1970-79," and "1980-89" were summed from single-year data to condense the presentation of these data. All numbers shown in boldface type were calculated by Eric M. Larson. Blank spaces represent zero. Note: In a letter dated January 9, 1997, NFA Branch Chief Nereida W. Levine stated that the "OTHER" category is "comprised of registrations where the form number is different from the other ones tabulated. An incorrect form number would be counted in that column." The National Firearms Act (NFA) was enacted in 1934, and became effective on July 26 of that year. All registrations ATF has reported for years before 1934 are errors. Some of ATF's documentation of this project was obtained from a Freedom of Information Act request. Because ATF provided no explanation, these documents are difficult to interpret precisely but I think that ATF's idea in providing them in a haphazard form was to confuse rather than be clear about errors in the NFRTR. ATF's effort did not completely succeed, however, because these materials unquestionably document errors in the NFRTR. Following this page, I have inserted four of these documents exactly as I received them. These documents raise troubling questions regarding the quality of data in the NFRTR. The reason is that these documents identify "our March 1995 project to clean-up (revise and amend) the entire National Registry." Really? Why would ATF initiate a project to "clean-up" the NFRTR after Mr. Miller, the Chief of the NFA Branch, had already attested to the NFRTR's accuracy? It is clear that data in the NFRTR were not nearly as accurate in 1995 as Mr. Miller alleges that they were in 1991, when he stated that "We have verified the accuracy of the information in the NFRTR." What happened from 1992 to 1996? Insert pages here In my judgement, Mr. Miller's statements regarding the accuracy of the NFRTR are not worthy of belief. The reason is that virtually all of the NFRTR data I have analyzed in this study are from the years 1934 to 1971. The "REGISTRATION ACTIVITY" represented by these data occurred from 25 to 62 years ago. If the NFRTR data were "accurate" in 1991, then, presumably, errors could only be detected for the years from 1992 to 1996. Yet, the data reported during 1992 to 1996 constitute very convincing evidence that there are significant numbers of misclassified and lost registration documents that occurred from 1934 to 1971 at least 25 years ago. Just what is going on here? How could these data possibly have been deemed "accurate" in 1991? Consider the first page, an "ERROR RESOLUTION REVIEW" which states that 151 documents were reviewed and that 22 percent (33 documents or, possibly, data fields represented in computer entries) contained errors of some type. Now turn to the next page. It is entitled "NFA's Monthly PRODUCTIVITY Report" that was "prepared 2/9/95" and contains the statement: "On approximately October 3, 1994, we began defining and separating the significant errors from the common errors." This document establishes that ATF has formally defined least two types of "errors" which exist in the current NFRTR data. One of these errors (e.g., "Approved form never updated in NFRTR") corresponds exactly to Mr. Busey's October 1995 statement about "a Form 4 waved in your face when you go in there to show that the guy does have a legally-registered Title 2 weapon." A "Form 4" is used by ATF to legally transfer the ownership of registered NFA firearms. If a transfer on a Form 4 is not updated in the NFRTR, the NFRTR will show that the firearm is still registered to its previous owner. Here, it appears, is proof that the type of error to which Mr. Busey was referring indeed exists. An "Approved form never updated in NFRTR" could certainly refer to a Form 4, and to other transfer forms. Consider that John D. LeaSure was initially convicted of nonregistration of NFA firearms. Also consider that a U.S. District Judge overturned these convictions on May 21, 1996, because of valid and reliable evidence that the NFRTR data are legally flawed beyond a reasonable doubt. This document establishes that an "Approved form never updated in the NFRTR" is an error that ATF has, in fact, encountered. The last two pages each apparently originated from the "National Firearms Act Branch," and are entitled "WEEKLY Quality Review Report." These last two documents are, to me, perhaps the most interesting from this group. The reason is that these documents show that somebody at ATF arbitrarily changed the definition of "Significant error" to "Error" in name only. Consider the document which bears the notation "Mon,6/19" indicating it was prepared in 1995, and which identifies six different "Significant errors": 1. Mispelled [sic] and/or Incomplete names. 2. Voided application--didn't indicate current firearm possessor. 3. $200/$5 remittance not posted. 4. Never mailed approved form to transferor 5. Approved wrong firearm to transferee. 6. Approved form never updated in NFRTR. This document also states: "Since 6/30/94 reviewed 25611 Errors 1567 Significant errors 373" and "Common Error rate .01% Significant error rate .01%" (underlines and boldface appear in the original document). The "Common Error rate" and the "Significant error rate" that I have quoted from this document appear to be incorrect. Specifically, 1,567 25,611 = .0611846, or 6.1 percent; while 373 25,611 = .0145641, which may be 1 percent or 1.4 percent or 1.5 percent depending on the criterion established for rounding the value up or down. The last document, which bears a notation "Mon,10/30" was, according to its day/date combination, also prepared in 1995. Note that the six "Significant errors" listed on the previous document have now been redefined simply as "Errors" in this document. The 36,903 documents (or, perhaps selected data fields) reviewed "since 6/30/94" reportedly contained 2,155 errors, producing an error rate of 6 percent. Viewed together, these last two documents indicate that sometime between June 1995 and October 1995, "Significant errors" were downgraded to and redefined as "Errors" by somebody at ATF. I suppose that one way to lower a "Serious error" rate is to simply redefine that term as an "Error" rate and that's apparently exactly what happened. Who is in charge of data quality control at ATF? Why was this decision made? What effect did it have on ATF's statements characterizing the accuracy of data in the NFRTR? 7. It appears, based on the preceding discussion and other evidence, that NFRTR data have the potential to be manipulated and may, in fact, already been illegally manipulated. In January 1996, some data on registered Thompson submachine guns were obtained from ATF, under a Freedom of Information Act request. These data were taken from the NFRTR, and used for research purposes. I have inserted 13 pages from the printout of this data run on the following pages. This data run apparently was made on December 22, 1995, because its pages bear that date. I have offered possible explanations of what the variables listed could mean in Table 9. For certain entries in some data fields, the meanings are to some extent clear (such as "RESTRICTED" indicating that a firearm is registered on Form 10, and that it cannot be transferred to a federally licensed firearms collector or other private citizen). This was a unique data run; consequently, all data gathered in the NFRTR may not be reflected, such as the "docket number" or other unique identifiers. Page inserts about here Table 10 about here There appear to be two lines per firearm in which information about the firearm is entered. The first line contains the serial number, barrel length, overall length, owner's name, address, and so forth. The second line contains other descriptive information about the firearm. Consider the entries which appear under the "SERIAL" data field on page 81 of the printout, which are intended to supplement information about the firearm described directly above: "VOID:", "F3:", "DUMMY CARD", "NO CARD". What is a DUMMY CARD doing in a data printout? What does NO CARD mean? What about VOID? Could unregistered firearms be quietly registered or re-registered in these data fields? One entry on page 57 of the computer printout reads: "DUMMY CARD EXCHANGED IN 1941 BY" with the rest of the entry censored. A number of records contain the statement "HAVE S/N CARD BUT NO DOCUMENT TO VERIFY FM# MAY BE DESTROYED OR ABANDONED" (pages 30, 34, 35, 54, 65, and 77). On page 52, we see the entry "NEXT UPDATE-TYPE OVER TOP LINE AS IS DUPLICATION." I can't help wonder why whoever wrote that entry didn't simply make that change then. What about "UNABLE TO RESOLVE DUAL"? (page 73). How good are these records? Consider page 35 of the computer printout: "FIREARM UNACCOUNTED FOR," and "F/A UNACCOUNTED FOR." Some firearms are "ABANDONED TO ATF" (see page 54); other entries state "F/A DESTROYED" (page 65). Are these firearms still counted as being "registered" in the NFRTR data that ATF has released? Table 10 Symbols on a Computer Printout of Selected 1995 Data from the National Firearms Registration and Transfer Record, and their Possible Meanings Symbol on printout Possible meaning of symbol SERIALserial number of firearm MFGmanufacturer of firearm (Colt, Remington, Ruger, etc.) TYtype of firearm (machinegun, Any Other Weapon, etc.) MODELmodel of firearm (i.e., Model 1928, Model 76, etc.) CALcaliber of firearm BLENbarrel length OLENoverall length of firearm SVCwhether a firearm is deactivated or serviceable LOSTwhether firearm or record of it was reported "lost" (?) DISPdisposition of the firearm, such as lost, destroyed, etc. MICROindicates that original document is microfilmed APPOTE. .unknown FORMform upon which firearm is registered (Form 4, Form 4467, etc.) COF . . .unknown NAMEcurrent owner's name CITYcity of current owner's residence STstate of current owner's residence Source: Bureau of Alcohol, Tobacco and Firearms. NFRTR data run on Thompson Submachine guns prepared December 22, 1995, under a Freedom of Information Act request. The possibility that computerized and other data can be illegally manipulated, or entered into any data system, or changed once in a data system, is probably a fact of modern life that all of us may just have to live with. But that is no reason for the inexcusably sloppy record- keeping that I have documented in this research, or for ATF not to be open about its firearms registration practices. Instead, ATF has decided to withhold from public release most of one of the few documents (the NFRTR data instruction manuals) that could possibly answer some of the questions I have raised. On March 29, 1997, I received 23 pages (some heavily censored) from ATF in response to my January 18, 1997, appeal of ATF's total denial of my Freedom of Information Act request for copies of these data manuals. Inspection of these 23 pages indicates that the NFRTR data base instruction manual is organized haphazardly, at best. There is no index, materials appear to have been randomly inserted, and the documents themselves are not of a uniform format. ATF's release of these 23 pages confirms what I believed all along that ATF's previous claim that these manuals are "related solely to the internal personnel rules and practices of an agency" is simply not true. Robert E. Sanders, Esq., the former ATF executive who talked with me about NFRTR certifications, said he was amazed at that claim: "I can't imagine ATF using that [reason] as a basis to withhold a document like that," he told me. "There's nothing in there [the NFRTR data instruction manual] that couldn't be released." In this study, I have made use of this new information to amend and clarify my research. But this new information does not address most of the questions I have raised (see NOTE at the end of this study regarding the NFRTR data manuals). The NFRTR data instruction manuals may contain descriptions of procedures that are to be used if, for example, the owner of a registered NFA firearm presents an ATF agent with a valid registration form that isn't in the NFRTR. It may also contain descriptions of the meaning of data and/or forms reflected in the column of "OTHER" data information about the registrations of 8,876 firearms. What is the legal basis under which these "OTHER" firearms were registered? Finally, the NFRTR data base instruction manuals may actually contain descriptions of procedures for re-using "docket numbers" or other identifiers, in such ways as to conceal lost or destroyed registrations, or even illegal registrations. It is Clearly Illegal for ATF to Claim That Issuing a Duplication Registration Document To a Lawful Owner is a Matter of ATF's Sole Discretion Under its current procedures at 27 C.F.R.,  179.142, ATF is not required to issue a duplicate NFA registration or transfer document. Instead, the regulation states that ATF "may issue a duplicate document upon such conditions as the circumstances warrant" (boldface added). I believe this regulation is illegal. But before discussing why, let's consider the products of its mischief. I've heard numerous tales about people who have said they found one or more NFA firearms in the estate of their father, uncle, or whomever, which ATF said were not registered, after these heirs sought to have the firearms transferred into their names. The only choice, said ATF, is to voluntarily surrender the firearm(s). Otherwise, ATF will confiscate them. Enraged, these people scoured their houses, other areas and in some cases they found original, valid paperwork. ATF then said it made a mistake; of course, the guns were registered. In other instances, ATF confiscated the firearms when the possessor could not produce a registration document. It is difficult for me to believe that any person would attempt to transfer the ownership of an NFA firearm, with full knowledge that the firearm was not registered, because of the legal consequences. An ATF agent stating that a firearm isn't registered can mean: (1) the firearm actually isn't registered; (2) ATF is using a false ploy of claiming nonregistration as a pretext for an informal, unwritten or unstated policy to simply confiscate the gun; or (3) that the firearm actually is registered, but ATF can't find its record of it. None of these are unreasonable assumptions. On the other hand, I've heard many accounts where ATF agents bent over backwards to courteously assist in the legal transfer of legally registered NFA firearms from the estates of deceased persons sometimes many years after the death occurred. The only problems I have heard about, in some instances, is when the firearm was transferred outside of the immediate family. This seems to mean that once such a firearm is no longer in possession of a legitimate heir, it is considered forfeitable. These firearms can be and definitely are confiscated. Obviously, a registration form is of critical importance. Of tremendous relevance here is an internal ATF Memorandum dated August 18, 1980 (bearing symbols CC-28,778 RMT and 22889), to the ATF Director regarding release of  6103 information, which states that a registration or transfer document is considered a "tax return." By law, any taxpayer is entitled to a copy of his or her tax return. Therefore, any lawful owner of an NFA firearm should be entitled to a copy of the registration document upon request. When hearings were held before the original National Firearms Act was enacted in 1934, the problem of the owner of a gun losing his or her registration form was discussed (see page 57 of the hearing, which I have inserted following this page). On April 18, 1934, Karl T. Frederick, President of the National Rifle Association of America, cautioned that the lack of ready access to a duplication registration would arbitrarily transform some law-abiding citizens into criminals. Insert page about here Consider today, April 8, 1997, exactly 10 days shy of exactly 62 years later Mr. Frederick's wise counsel: "I think the owner ought to be able to get one if it is lost," he said. "I think that machinery ought to be made simple. If not, in the actual operation, you are going to create criminals." Mr. Frederick's proposal made sense, and it should have been enacted. Nothing of the sort, however, was enacted under the original NFA. Nor was any such provision enacted when the Congress extensively revised the NFA under Title II of the Gun Control Act of 1968. What happens, then, when an original registration is lost or destroyed in the course of normal human events, or natural disasters? Today, under 27 C.F.R.  179.142, ATF states that issuing a duplication registration is a matter of ATF's sole discretion. The person who lost the document, or his or her lawful heir, is totally at the mercy of ATF, because there is no requirement that ATF issue a duplicate registration or transfer document. If persons requesting a duplicate document from ATF can't know if the registration or transfer document was ever filed (such as an heir or executor of the estate) they can't know if ATF can't find the registration, or if the gun was ever registered, or if ATF is flat out lying to them as a pretext to confiscate the firearm. Incredibly, in my opinion, instead of creating regulations to deal sensibly with, as we have seen, "the condition where people who registered firearms under the original National Firearms Act at age 65 would now be 112 years old," ATF has written a regulation which leaves the issuance of a duplicate registration as a matter for ATF's sole discretion. I believe this regulation is illegal, because it potentially unilaterally denies the right of a taxpayer to obtain a copy of his or her tax return (registration document) on demand with no means to appeal. The Tax Code, on the other hand, requires that a taxpayer must be given a copy of his or her tax return upon request. This issue isn't even debatable. Consequently, in my judgement, the law and common sense requires ATF to immediately revise 27 C.F.R.  179.142 as follows: When any fForms 1, 2, 3, 4, 5, 6A, or 10 (Firearms), or document that the Director has issued or approved evidencing possession of a firearm is stolen, lost, or destroyed, the person losing possession (or, if deceased, his lawful heir, personal representative, or the executor of his estate) will immediately upon discovery of the theft, loss, or destruction report the matter to the Director. The report will show in detail the circumstances of the theft, loss, or destruction and will include all known facts which may serve to identify the document and the firearm. Upon receipt of the report, the Director will make such investigation as appears appropriate and shall, within 60 days, may issue a duplicate document or form, or a certification, under penalty of perjury, that such firearm is not registered in the National Firearms Registration and Transfer Record, unless it can be conclusively proven that the firearm identified by the form or document has been destroyed upon such conditions as the circumstances warrant. I have included the perjury aspect because, under Title 26, U.S.C.,  5861(l), it is illegal for any person (including an ATF agent or employee) "to make, or cause the making of, a false entry on any application, return, or record required by [the NFA], knowing such entry to be false." ATF operates under the presumption that citizens must comply with the law. Citizens must operate under the presumption that ATF must comply with the law. Why Haven't More People Spoken Out? The only protection that the lawful owner of a registered NFA firearm has is his or her copy of the registration or transfer document which, as we have seen, is prepared in duplicate original. If ATF later takes a position (such as that under Title 26, U.S.C.,  5851(b)) that the owner has received or possesses "a firearm transferred to him in violation of the provisions of this chapter," and moves to confiscate it, the lawful owner has no means to challenge ATF's record-keeping in the National Firearms Registration and Transfer Record, since none but ATF personnel have so far been allowed to inspect these records. As we have seen, there is more than reasonable doubt that these records are seriously flawed, and ATF personnel have vested interests in their representations of the accuracy and integrity of these data. Indeed, on May 21, 1996, a federal court overturned five convictions for nonregistration because the NFRTR data were not accurate beyond a reasonable legal doubt. Why don't more people don't speak out? First, it is necessary to understand that relatively few people own registered NFA firearms. All those people I've ever known have been extra careful to make sure their guns are legal; it's not a topic I've seen any of them take lightly. So, one answer is that this group of firearms owners is relatively small, and so there's not millions and millions of people involved. Another answer is that most people I've encountered who lawfully own NFA firearms are pretty reticent. As a group, I've found, most of them are reluctant to let it be known that they own such firearms. There are many reasons, I think, for this: fear of being robbed; or a possible social stigma from people who don't understand or appreciate historical artifacts; or a desire to keep a relatively low profile to avoid any further tightening of the already-strict regulations regarding the possession of these firearms. More fundamentally, the thought of being jailed or losing a significant amount of money in a fight with ATF over a forfeiture action (which happens when ATF decides that you don't legally own an NFA weapon that you believe you legally possess). I don't personally believe that ATF agents typically "rough up" the owners of registered NFA firearms, because they don't need to. The job of a law enforcement agent is to fight crime in various ways, and that includes scaring people by deception, tone of voice, and evasion to try and make them fearful criminals are used to this; ordinary, law-abiding citizens are not. Everyday actions like commanding suspects (i.e., "Assume the position!"), patting them down for weapons, sticking guns in their faces and threatening to kill them, handcuffing them, and publicly hauling them off in front of newspaper reporters who often have cameras (termed "the perp walk," short for "perpetrator," again part of the specialized argot of the underworld) to be jailed, arraigned, or interrogated by an expert are terrifying to most people who are unfamiliar with law enforcement. In the hands of an expert seemingly harmless statements, and unintentional contradictions, can often be used with great success against persons unaware of the analytical framework to which they are being subjected. Another reason is confusion about the law, fear of ATF as an arm of the Internal Revenue Service, and a lot of nervousness about lost paperwork. Consider the following statement which ATF could legally (citations of law or regulation are listed in the Endnotes) make: We have received your report to the ATF Director of a lost or destroyed Form 4467 registration for a Model 1928 (Navy type) Thompson .45 caliber Submachine gun, which you say you lawfully inherited from your father when he died last month, and which you claim is a valuable collector's item valued at more than $10,000 because of its condition and historic association with U.S. Armed Forces in Nicaragua prior to World War II. You state that your father registered this weapon during the 1968 amnesty period, that you have seen the registration document that you describe, but you also say that you are unable to locate it. As you know, the NFA requires the owner of any NFA weapon to keep the registration document as proof of lawful ownership, and for the owner to show it to us any time we ask to see it. We also note that if you don't have your registration document and the weapon involved is a machinegun, we don't need to prove that the weapon isn't registered for a Court to convict you of unlawfully possessing an unregistered machinegun. The law provides that our registration records aren't open to inspection, so you and the Court have to take our word for what's in our files even if we decide to prosecute you. Under the regulations we have established, we have decided that we may give you another copy of the registration document that you say you are entitled to, after we do an investigation to see if we like your story. In the meantime, an ATF Special Agent will personally come to your home to take physical custody of this weapon within the next 48 hours, since you have stated that it is not currently registered to you. We also will notify law enforcement officials of your state that you have possessed this weapon and that we have confiscated it. Our computer files disclose that you currently have a valid federal collector's license, which entitles you to acquire firearms that we recognize as being defined as "curios or relics." We now inform you that we must inspect all of your firearms and entries in your bound volume of records, either where you keep the firearms or at one of our local offices that we direct you bring them to, and that we can do this once per year anyway without any reasonable cause or warrant. You should have nothing to fear from us. All we seek to do is determine whether or not we think you are a legitimate firearms collector. After we have inspected your firearms and interviewed you, a decision will be made regarding your request for a duplicate copy of a registration document for a weapon that you say you lawfully inherited. We are the only people who can issue you a duplicate copy of a registration document, as you loyally recognize. A person unfamiliar with firearms laws might regard the fictitious statement I have just written as mighty funny but I can't think of a single collector of "curio or relic" firearms who would laugh at it. Under the regulations ATF has written, the lawful owner of or heir to a registered NFA firearm is not entitled to a duplicate copy of his or her registration yet, the law clearly provides that he or she can be prosecuted for losing it (not possessing it). People also, I believe, don't speak out because of the lack of access to valid and reliable information, and some instances where ATF has deliberately and clearly made incorrect legal decisions in its administration of the law. As I have tried to demonstrate in this article, and other research (such as my 1996 testimony before the House Appropriations Subcommittee), the NFA is very complex. Most people don't understand certain aspects of the law, or what their rights are, and ATF can at times not be a user-friendly organization. As an intelligence agency and a law enforcement agency, ATF exudes some of the mysteries of tradecraft (using digraphs and other symbols in the headings of its letters to people), and its institutional posture is mainly one of (1) retaining all possible discretion in its actions; (2) refusing to release certain information that would be useful to people who are trying to figure out why ATF has apparently made legally incorrect decisions; and (3) at times, refusing to discipline its employees. For example, no ATF employee who threw away or destroyed NFA registration documents, as documented in the LeaSure case, was fired or disciplined. In fact, ATF allowed them to continue doing the same work. Finally, being treated fairly at the hands of ATF may come very at a very high economic price. Legally contesting the possibly unlawful seizure of a firearm runs about $25,000 and up. A forfeiture action is expensive; it means posting a bond, and certain time deadlines have to be met. For a gun that may be worth as much as $1,000 to $5,000, it still is cheaper for the owner to simply surrender ("voluntarily abandon the weapon to the Bureau," in ATF language) the firearm. There are instances where this has definitely happened. For several years I have published my address and telephone number along with articles I've written, as well as in firearms reference books. I've found that a significant number of people want information, and in some instances the exchanges have been mutually educational. I've also been struck by how many grown men seem absolutely terrified of ATF, the number of times I've been asked "Please don't ever mention my name," and the people who profess to be outraged yet refuse to become personally involved. Where Do We Go From Here? ATF may choose to admit to the errors that I have identified in this research and voluntarily establish another amnesty period to correct them, but I doubt it. The reason is that in my personal experience, ATF's institutional culture would oppose another amnesty period, and would go to extreme lengths to try and disprove any challenge to the validity and reliability of data in the NFRTR. I also believe that ATF may simply stonewall any inquiries, or release very limited information that is essentially worthless. The other day, I saw a list which ATF might well put to use in the near future regarding errors in, and the integrity and accuracy of, the NFRTR: 1. Never admit doing anything wrong. 2. If unethical activities are discovered, your staff did it without your knowledge. 3. If found breaking the law, insist it is an ethics issue. 4. If that doesn't work, insist the other party has been doing it for years. 5. If that doesn't work, insist on rewriting the law so you won't do it again. 6. If that doesn't work, interpret the law to exempt the Executive Branch. 7. If that doesn't work, hide behind the skirt of the Attorney General. Alternatively, the Department of Justice or the Congress may choose to become involved; either entity could move the action along considerably quicker than a private citizen, particularly by using a forensic audit of the NFRTR as an investigative tool. In my judgement, it is may be somewhat doubtful that ATF would try and change any NFRTR records to try and conceal whether its personnel have added lawfully registered firearms to the database, when confronted with a valid registration document, or may have registered or transferred firearms illegally by mistake or otherwise. The reasons are that (1) the chain of evidence would be destroyed, compromising the entire NFRTR; (2) changing records such as the "docket number" or its equivalent to reflect a different time period, would constitute destruction of evidence and obstruction of justice; (3) serious legal penalties (see Title 26, United States Code,  5861(l) and  5871) would apply to any person who knowingly made or caused the making of a false entry on any NFA record. These penalties include a fine of up to $10,000 and up to 10 years imprisonment for each violation; and (4) if a coverup of this sort were attempted, somebody at ATF would talk. As was the case with the Busey matter, there are honest public servants at ATF who would not tolerate the destruction of evidence when presented with a lawful inquiry or go along with a coverup, or an attempted coverup, in a situation of this gravity. It seems clear, however, from the evidence that I have presented, that obtaining complete and truthful answers from some ATF personnel to questions about their personal knowledge about the accuracy and integrity of the NFRTR, will probably require that formal, sworn depositions or testimony from them be made under oath, under penalty of perjury. I believe that there is a potential for some ATF employees to commit perjury, because I have detected this and documented it in the past, using documents that ATF itself provided. The Freedom of Information Act, I have found, can work. During one of the House Appropriations Subcommittee hearings on ATF last year, then-Chairman Jim Lightfoot stated: I would say that ATF is a bit like its hero and most famous agent, Elliot Ness: a bit of a throwback to an earlier era, with a lot of talent and heart, but having trouble following the new, stricter law enforcement rules of the 1990s. ATF is a troubled agency, but not, in my opinion, one that is out of control On the other hand, there are way too many questions about ATF operations that we cannot ignore: mismanagement in the use of funds, lying on search warrants, managers being rewarded despite gross incompetence, harassment of firearms dealers, and unchecked racism and sexism. Quite frankly, I am concerned about a pattern at ATF of not disciplining personnel, particularly the managers. The two ATF field commanders at Waco still work for the Agency, despite the fact that a Treasury report found that they had made serious errors in judgement. I want to make it absolutely clear that I refuse to insult the memory of agents who have died in the line of duty by tolerating lazy, incompetent, and irresponsible employees. And I will do whatever needs to be done to increase outside oversight personnel actions, strengthen the Treasury Inspector General, and even take all the ATF agents out of the jurisdiction of the Merit Systems Protection Board to get these kinds of people out of ATF. I don't know the extent, if any, to which Mr. Lightfoot's remarks represent the current thinking of the Subcommittee's present Chairman, Jim Kolbe, but that's not why I cited them. The reason I cited Mr. Lightfoot's remarks is my belief that any resolutions of problems that may exist with ATF involve, ultimately, political processes that are handled within the government. An individual person, or even an interest group, is simply not capable of action on that scale. The Congress and the Executive Branch often see things very differently their eyes view the world through very different lenses, as it were. Conflict is inherent in articulating these viewpoints, by definition most legal processes are adversarial, and this is the way that our government has been structured. Often, neither side has a particularly easy go of it and to some extent each side must give and take in mutual compromise. Sometimes, interested individuals at a Public Witness Hearing can provide information that each side may find useful in debating issues of public interest. True government by the people isn't a spectator sport. NOTE: ATF's recent partial release of certain NFRTR data manual documents was accompanied by a cover letter which, like many letters I have received from ATF, answers questions or issues that I did not ask or raise. Although I stated in my original Freedom of Information Act petition, and my appeal of its total denial, that I was not seeing any confidential information and only sought information describing how confidential information was actually entered into a data system, ATF nevertheless stated that I was seeking records that contain information about individuals ("You have not articulated any public interest in releasing the documents, and your personal interest in obtaining disclosure is not relevant"). My original request, and my appeal, stated that I was not seeking any confidential information only information about how that information was entered into a data system. My appeal also stated, in part, that I was seeking the NFRTR data manuals because the very real possibility that firearms which are legally registered are not recorded in the NFRTR, ha[s] serious criminal consequences. Specifically, a violation of the NFA for possession of a firearm not registered in the NFRTR may incur a fine of up to $250,000 and up to 10 years imprisonment. In the interest of fair and equal justice, and due process, I believe that the public interest requires ATF to disclose the information that I have requested; (4) I believe that the documentation I am requesting would assist Congressional decision making regarding an independent audit of the NFRTR by a qualified outside entity with staff who possess the necessary security clearances; and (5) this documentation may assist in determining whether ATF has added firearms to the NFRTR for which it previously had no record, but was informed of their existence by a lawful owner producing a legally valid registration. There are some unlikely outcomes involved with certain FOIA requests. Asking ATF to "send me documentation of how your employees illegally registered NFA firearms on Form 4467 during 1969 to 1971, and since 1971" would probably result in a letter stating that no such documents could be located, which may be true. Some things aren't written down, but that doesn't mean they don't happen. Remember what Mr. Schaible answered when I asked him if ATF had ever added an NFA firearm to the NFRTR because a lawful owner produced a valid registration document, and ATF had no record of it: "Yes. I assume that's happened." Finally, the NFRTR data manual information that ATF sent me didn't contain any explanation of what "OTHER" category "REGISTRATION ACTIVITY" meant. NFA Branch Chief Nereida W. Levine told me that these data correspond to identifiable forms and form numbers. How can the name and number of a government form used to register NFA firearms be classified? If it is classified, why is it classified? I invite readers and other interested persons to send me information which may be of legal value in resolving questions about the accuracy of the NFRTR. I can be contacted at: Eric M. Larson, P.O. Box 5497, Takoma Park, Maryland 20913, telephone (301) 270-3450. As was my practice in researching my booklet, Variations of the Smooth Bore H&R Handy-Gun, I will respect all requests for anonymity. Endnotes