1. See, generally, Consumer Product Safety Commission, Directorate for Hazard Identification and Analysis, "Unvented Gas Space Heaters: HIA Hazard Analysis Report," July 1978.

2. The amount of ventilation depends on the size of the heater and the rate of air infiltration. With a 40,000 BTU heater, for example, the largest allowed under Z21.11.2, a standard-sized window should be opened at least two inches to ensure sufficient ventilation.

3. See, for example, "Minutes of Meeting of Subcommittee on Standards for Unvented Gas-fired Space Heating Appliances" (Cleveland, November 10-11, 1970), item 6, p. 6 (cases involving clothing ignition cited in letter forwarded by the secretary of commerce); "Minutes of Meeting of Subcommittee on Approval Requirements for Unvented Gas-fired Heating Appliances" (Cleveland, September 11-12, 1962), item 5, p. 19 (newspaper clipping sent by NFPA about fire incident involving an open-front room heater).

4. The CPSC list, now in general disrepute, purported to combine the frequency and severity of reported injuries into a single index. A major criticism, aside from problems with the quality of the data collection, was the failure to factor out injuries that could not possibly be reduced through regulation. Stairs, for example, topped the list for many years. Five of the next seven products were sports-related.

5. CPSC, "Unvented Gas Space Heaters: HIA Hazard Analysis Report," p. 12.

6. The most likely explanation is that between 1974 and 1975 the CPSC received seventy death certificates "attributed to carbon monoxide emissions from gas space heaters," but these covered an eighteen-month period and did not distinguish between vented and unvented gas space heaters. Based on seven reported fatalities related to unvented heaters, the CPSC estimated a national total of fifty-six in 1977. This extrapolation is subject to the same well-known criticisms of CPSC's NEISS system.

7. Max Singer, "The Vitality of Mythical Numbers," The Public Interest 23 (Spring 1971). See also Peter Reuter, "The (Continued) Vitality of Mythical Numbers," The Public Interest 75 (Spring 1984).

8. The CABO (Conference of American Building Officials) One and Two Family Dwelling Code provides, for example, that "only unvented fuel burning heaters equipped with oxygen depletion sensors and listed in accordance with ANSI Standard 'Z21.11.1' 1978 shall he permitted."

9. UL, which began by testing electric appliances, now has six departments: Electrical; Burglary Protection and Signaling; Casualty and Chemical Hazards; Fire Protection; Heating, Air Conditioning, and Refrigeration; and Marine. A few of UL's standards cover gas appliances, overlapping directly with some of AGA's standards. This can create a serious problem if the competition is not just in price, as is the case when smaller labs offer to certify products to UL standards. In these rare instances, the laboratories "compete" on safety as well. Such competition tends to reduce safety requirements.

10. Because of the extensive review and comment procedures utilized by AGA, most standards are submitted to ANSI as "uncontested cases." For example, Z21.11.2 was approved by the Board of Standards review on October 12, 1984, without discussion.

11. American Gas Association, "Procedures for Accredited Standards Committee Z21 on Performance and Installation of Gas Burning Appliances and Related Accessories" (approved by ANSI on July 18, 1984; mimeographed, n.d.).

12. See, for example, "A Resume of the Voluntary Standards Development Procedures for Z21, Z83, and Z223 Standards" (AGA Labs, July 1981; mimeographed), 5.

13. American Gas Association, American Standard Approval Requirements for Gas-fired Room Heaters. Volume II, Unvented Room Heaters, 14th ed. (New York: American Gas Association, 1962), sec. 1.5.4, p. 3 (ANSI Z21.11.2-1962; UDC 644.1:696.2).

14. Ibid., sec. 1.6.3.

15. Ibid., sec. 1.7.8, p. 4.

16. Ibid., sec. 1.9.4.

17. AGA Labs, "Minutes of Meeting of Subcommittee on Approval Requirements for Unvented Gas-fired Heating Appliances" (Cleveland, December 8, 1964), item 5, p. 3.

18. AGA, Unvented Room Heaters, sec. 2.15, p. 15.

19. AGA Labs, "Minutes of Meeting of Subcommittee on Approval Requirements for Unvented Gas-fired Heating Appliances" (Cleveland, January 11, 1962), item 8, pp. 6-7.

20. For information about the proposal to eliminate these standards, see John Herbers, "U.S. Plans to End Nationwide Code for Construction," New York Times, August 18, 1985, 1.

21. Philadelphia Electric Gas System, "Unvented Room Heater and Decorative Indoor Appliance Installation Standard: Final Report" (photocopy, n.d.; attached as appendix to American Gas Association Labs, "Compilation of Comments Received on Proposed Revisions to American National Standard for Gas-fired Room Heaters, Volume II, Unvented Room Heaters," December 1975), 3.

22. American Gas Association, American National Standard for Gas-fired Room Heaters. Volume II, Unvented Room Heaters, 19th ed. (Cleveland: American Gas Association Laboratories, 1983), preface, p. i (ANSI Z21.11.2-1983).

23. Other examples from the case studies of the unusual interaction between installation codes and product standards include wall pass-through systems for woodstoves and flexible nozzles for aviation fire extinguishers.

24. American Gas Association Laboratories, "Compilation of Comments Received on Proposed Revisions to American National Standard for Gas-fired Room Heaters," September 1979, p. 2 (R. C. Swangler, The Peoples Natural Gas Company).

25. The severity of burn injuries is a function of both the surface temperature and the time of exposure (or the "response time" before ending the exposure). According to a report prepared for the CPSC, contact with a bare metal surface 140° hotter than room temperature will not produce a third-degree burn unless contact is maintained for at least twenty seconds. Therefore, by lowering surface temperatures, a longer response time is necessary for the same type of injury to result from contact with the heater.

26. Bond paper is nonporous and, therefore, unlikely to ignite under the old test method, even though a similar exposure might actually ignite clothing. Minutes of the subcommittee indicate that someone argued as early as 1963 that "it would be desirable to use a material of greater porosity than bond paper so that oxygen could get to the charring surface." Selecting an appropriate level of porosity for test materials is one of those seemingly "technical" decisions that has significant real-world effects and often cannot be answered scientifically. Porosity is really a correlate of safety. Bond paper, for example, lacks the porosity to reflect the hazards of clothing ignition. Cheesecloth, on the other hand, may be too porous. "The air goes right through it," argues a representative of Consumers Union. By using muslin instead of cheesecloth, CU concluded that an electric space heater approved by UL (using cheesecloth) was unsafe.

27. Z21.11.2 did not ban the model by name, but by features—that is, by requiring such features as automatic ignition.

28. Federal Register 40 (September 5, 1975): 4112.

29. In industry's eyes, an unvented heater must be so by design; "vented" heaters installed without venting are nevertheless "vented." The CPSC wasn't sure. Injury data often do not include reliable information on venting, causing considerable disagreements in data interpretation. In a study of 533 death certificates related to gas-fired heaters, for example, the CPSC classified 72 as unvented, 82 as vented, 64 as "judged to be unvented," and 277 as "venting condition not specified." The remaining 38 were located in campers, vans, or trailers. See CPSC, "Unvented Gas Space Heaters: HIA Hazard Analysis Report," 15. A trade association representative familiar with the CPSC "in-depth investigations" claims that many of the injuries categorized as "unvented" heater cases actually involved vented heaters that were not properly vented.

30. One reason is the limited background and training of CPSC personnel in accident investigation. The agency has since improved its training program and increased its use of expert consultant to conduct investigations.

31. Federal Register 42 (September 14, 1977): 46072 (notice that it is not necessary at this time to address any hazards that may be associated with vented gas-fired space heaters); Federal Register 42 (February 14, 1978): 6253 (proposed ban on unvented gas-fired space heaters).

32. The unvented heater was one of sixteen products identified by the National Commission on Product Safety as warranting federal standards. Naturally, this created an impression among many in industry that the CPSC was predisposed to regulate the gas heater. See National Commission on Product Safety, Final Report Presented to the President and Congress (Washington, D.C.: Government Printing Office, 1970), 12-34.

33. For a detailed description of the devices, see Charles S. Lamar, "Oxygen Depletion Sensor Improves Safety of Gas-fired Heating Equipment," Appliance Engineer 5, no. 1 (1971): 21-28.

34. AGA Labs, "Minutes of Meeting of Subcommittee for Unvented Gas-fired Heating Appliances," Cleveland, January 20, 1972, item 6, p. 2.

35. Consumer Product Safety Commission, "Briefing Paper: Proposal to Ban Unvented Gas-fired Space Heaters" (prepared by the Fire and Thermal Burns Program Staff, CPSC, December 21, 1977), p. 3.

36. Firms are so worried about grading safety that they generally refuse to have their gas appliances tested for efficiency—something required by other laws—at the same laboratories that test their products for safety, even though doing so would certainly be less expensive.

37. Esher Kweller, National Bureau of Standards, "Response to Request for Support: Unvented Space Heater Project" (Memorandum for Walter Leight, through Andrew Fowell, dated January 16, 1979), 2-5.

38. Letter from Esher Kweller, National Bureau of Standards, to Stanley Blachman, chief methods engineer, American Gas Association Laboratories, March 16, 1979, attachment F, an excerpt from American Conference of Government Industrial Hygienists, Documentation of the Treshold Limit Values for Substances in Work Room Air, 3d ed., 1971.

39. Memorandum from Leo T. Duffy, M.D., to George Anikis, Office of Program Management, CPSC, October 3, 1977.

40. Memorandum from Dr. Peter Preuss to Bert Simson, Office of Program Management, CPSC, July 3, 1979, p. 2. Concern for health effects prompted Esher Kweller of the Product Performance Engineering Division of the National Bureau of Standards to argue for a 19 percent shutoff level. (As explained earlier, a higher shutoff level for oxygen means that the heater will shut off at a lower carbon monoxide level.) See Kweller, letter to Blachman, March 16, 1979, p. 2.

41. Letter from Frank E. Hodgon, AGA senior vice president for consumer affairs and safety, to George Anikis, Office of Program Management, CPSC, July 26, 1978, p. 3.

42. American Gas Association Laboratories, "Notice to Manufacturers of Unvented Room Heaters," August 9, 1979.

43. Federal Register 45 (January 18, 1980): 3762 (proposed rule).

44. CPSC, "Briefing Paper: Proposal to Ban Unvented Gas-Fired Space Heaters," 3.

45. See, for example, the following discussion of labeling in the CPSC's Federal Register notice: "Several commenters still prefer the ANSI label wording which specifies a minimum 'fresh air opening' depending on the output size of the heater rather than the proposed label's suggestion to 'open a window an inch or two for fresh air.' ... The Commission considers that the proposed label's wording ... is clearer and more readily understood than the ANSI label" (Federal Register 45 [September 17, 1980]: 61918).

46. Petition for Exemption, in the Matter of Petition of City of Victorville, State of California, for Exemption for 15 U.S.C. 2075(a) as It Applies to Commission Safety Standards Requiring Oxygen Depletion Safety Shutoff Systems (ODS) for Unvented Space Heaters (16 CFR Part 1212), April 16, 1982, p. 4 (CPSC Application no. SH 82-12).

47. The CPSC staff disagreed about whether the standard even had preemptive effect in the first place. See "Briefing Package on Petition SH 82-1 through SH 82-23: Requesting Exemption from Preemption by the Safety Standard Requiring Oxygen Depletion Safety Shutoff Systems (ODS) for Unvented Gas-fired Space Heaters; 16 C.F.R. Part 1212" (transmitted to the commission by memorandum from Douglas Noble, Office of Program Management, January 19, 1983).

48. Letter from Joe Beck, Kilpatrick & Cody, representing Atlanta Stove Works, to Nancy Steorts, CPSC chair, April 22, 1983.

49. Dissenting Opinion of Vice Chairman Saundra Brown Armstrong, in the Matter of Revocation of the Mandatory Standard for Unvented Gas-fired Space Heaters, 16 CFR 1212.

50. In his critical examination of five CPSC regulations, W. Kip Viscusi concluded that "more than any other standard ... the standard for unvented gas-fired space heaters may be desirable (Regulating Consumer Product Safety, 99).

51. The clear trend in legal decisions affecting consumer products is toward strict liability, minimizing the role of any negligence on the part of the consumer. But that depends on the product being considered unreasonably dangerous. Unless the gas space heater was considered defective for failing to have an ODS device—a line of argument that has been tried unsuccessfully with cars not equipped with airbags—then fatalities caused by carbon monoxide would not be considered the manufacturer's fault.

52. Space heaters equipped with an ODS would have to be considered "unreasonably dangerous" in order for liability to attach. Since this phrase has been stretched to include a variety of "reasonably foreseeable" mishaps and failures, however, the fear is not misplaced.