Of the estimated 1,000 patrons inside the Cocoanut Grove nightclub when the fire occured in 1942, 492 died.1 According to “Searching for Answers” (NFPA Journal, May/June 2000): “Developments in fire science may shed light on the 57-year-old mystery surrounding the Cocoanut Grove fire that killed 492 people.” I maintain that there is no mystery.

The cause of the terrible loss of lives has been published for a number of years. Casey C. Grant wrote an article on this fire in the May/June 1991 issue of NFPA Fire Journal. My letter to the editor, published in a subsequent issue, fully explained that the deadly glued-up combustible acoustical tile was the cause of the massive loss of life. A follow-up letter from a reader told of the high flammability of the adhesive globs that had been left on the ceiling for years after the tiles had been removed.

Proof of My Hypothesis

In 1980, I learned that the New York Times had published a bound copy of its pages covering major disasters, including the Cocoanut Grove fire. One photograph showed the globs of adhesive, proving that the combustible acoustical tile had been glued to the ceiling. I immediately related this information to the Hartford Hospital fire of 1961, in which glued-up acoustical tile caused multiple deaths.

Some History of the Combustible Fiberboard Menace

After deadly fires in a soldier’s hostel in Newfoundland and in Mercy Hospital in Iowa, fire officials, in the late 1940s, grew to realize that there was a new problem-the use of combustible fiberboard as interior finish. In April 1949, a disastrous fire occurred in St. Anthony’s Hospital in Effingham, Illinois. Seventy-four died, principally infants in the nursery and the nuns and nurses who would not leave them. The hospital had been finished with combustible tile.2

I can personally attest that the industry vigorously fought any attempts at regulation, which were characterized as “taking the bread out of the mouths of Louisiana cane farmers.”3 The National Fire Protection Association (NFPA) was threatened with lawsuits. The industry had de-veloped a U.S. De-partment of Commerce-accepted (but never validated) “in-dustry standard” fire test,4 which amounted to a fire in a thimbleful of alcohol. A light flame-retardant coating enabled combustible fiberboard tiles to pass this “test.”

A change in industry attitude came about not from reasoned explanations by fire protection experts but as the result of a lawsuit. The management of St. Anthony’s Hospital obtained a large sealed settlement against the manufacturer of the combustible fiberboard. I was told that the Sisters had been told that the interior was a hazard. They wrote to the manufacturer, who assured them that it was not and cited the test mentioned above.

The manufacturer’s management appraised the potential for similar massive liability judgments or settlements and directed the officials to cooperate in developing standards as a matter of self-protection. Thereafter, the opposition of the industry gave way to cooperation in developing fire standards. It is just about impossible to purchase ceiling tile today without being made aware of its flame-spread rating.

Through the years, square miles of acoustical tile were installed over old deteriorated plaster. Either a wooden structure was erected below the ceiling or the tile was glued to the ceiling.

The Hartford Hospital fire of 1961 clearly showed the special hazard of the adhesive used to glue up the tile. The corridor ceilings were of combustible acoustical tile glued to a gypsum-board suspended ceiling, which concealed the piping. The hazard had been recognized, and the tiles were painted with a flame-retardant coating. A fire that killed nine patients developed in a soiled linen chute. It rolled out and attacked the ceiling. The fire roared down the corridor.

Samples of the tile were tested and showed that the reduction in flame spread potential was not sufficient to give anywhere near the protection needed for a fire this intense. An entire 25-foot section was removed and sent to Underwriters Laboratories (UL). The flame spread was very high, caused by the adhesive that attached the tiles to the ceiling.5

The Clark County Fire Department report on the MGM Grand Hotel fire in Las Vegas in 1980 notes that 12 tons of adhesive were used to attach the tiles to the casino’s ceiling. The flammable adhesive added a large load to the fire, described as moving faster than a man could run.6

A most serious related hazard is often created when a building is remodeled. The code requires that a new ceiling that meets flame-spread requirements be installed. No code of which I am aware requires that the old ceiling be removed.7 The new “fire-rated” ceiling is installed below the old ceiling, leaving the dangerous combustible ceilings above. Fire will burst down out of that void. Two firefighters died in Wyoming, Michigan, when fire burst out of the ceiling. Even then, the city did not amend the code. Sixteen persons died in the fire at the John Sevier Nursing Home in Johnson City, Tennessee. The fire involved combustible tile left in place when a new grid ceiling was suspended below it (see Figure 1).8

The hazard of combustible fiberboard products is still imperfectly understood, particularly by some fire suppression forces. At times, it is regarded as a problem for “fire prevention,” particularly where the disgraceful “Iron Curtain” between fire prevention and fire suppression exists. It is also not necessarily understood by architects, particularly when existing buildings are concerned. Grandfather clauses permit known hazards to continue in existing buildings. Although it is legal, it often is considered “unfair” to require the removal of a once-approved material.

I found glued-up acoustical tile in a dormitory at a well known southern university. The university was quite concerned and agreed the tile should be removed. It said it had employed an architect to bring the building “up to code.” The architect either didn’t recognize the hazard or, if it was recognized, did not report it because it was grandfathered and thus legal.

In 1982, two Boulder, Colorado, firefighters died in a training fire in a building lined with combustible fiberboard. An article noted, “The building was toured by many people prior to the drills, and no one noted any problem with the interior finish of low-density fiberboard.”9

In Ol’ Professor, January 1999, I noted a National Institute of Occupational Safety and Health (NIOSH) report of two firefighter fatalities that occurred when a basement ceiling was pulled. A burst of fire from the old combustible tile ceiling engulfed the area. Two Washington, D.C., firefighters died on the first floor of a house as the result of an unexpected burst of heat from a fire in a finished basement. The basement ceiling was combustible tile on wood trusses. A truss void or “truss loft” provides the space for carbon monoxide to accumulate and burst out. Collapse is not the only hazard posed by trusses.


Factory Mutual has expressed concerns relative to steel sandwich panels filled with expanded polystyrene (EPS).10 These panels burn furiously and would require a special line of wet pipe sprinklers along the wall. Alternatively, FM requires an approved gypsum board thermal barrier. This also applies to metal deck roofs with EPS barriers. I must say that the thermal barriers may pass laboratory tests, but the reliability of gypsum board in place is questionable because of how easily it can be damaged or tampered with by uninformed employees. I would be wary of entering any such area. Use your thermal imaging camera to determine if fire is behind the gypsum board.

Steve Corwin of Oregon sent me an article from Fire Research News, a British publication, about large insulated sandwich panels (LISPs) in buildings-metal sandwiches filled variously with expanded polystyrene, polyurethane, or mineral fiber (noncombustible.) Two firefighters lost their lives in a fire in a chicken processing plant in England that had such panels. The panels are widely used in lightweight steel structures. The hazard is not apparent to the casual observer because the filling is invisible. Reports of fires involving these panels universally related the presence of large quantities of black smoke and some collapses of the panels.

If you have such steel buildings in your area, ask the management about the panels’ contents. The United Kingdom home office recommends diligent attention to housekeeping and the integrity of exits for occupants. A fire in the building could quickly involve the panels and generate so much smoke that occupants might not be able to escape. Check for further developments.


Piolite Plastics Co. offers UL-listed and FM-approved lightweight plastic panels to be installed on a hanging ceiling below sprinklers to hide them and yet not interfere with sprinkler operation. Such panels are not new. Years ago, lightweight panels were installed at Oak Ridge National Laboratory. When a door was opened, the panels would fly up and come down in disarray. A maintenance man used his “common sense.” He laid plywood all over the ceiling to hold the panels down, thus defeating their purpose.

Every time I hear “common sense” used in discussing a fire problem, I get upset. Common sense is important when it is drawn on in age-old interpersonal relationships. Fire, however, is a technical subject and demands knowledge of the problem and the ability to draw the best solution from a number of choices. In fire situations, common sense is like tossing a coin. It has only a chance of being correct.

Permit me to repeat myself: When sprinklers are ordered to be retrofitted, the owner often objects because of “excessive cost.” Often a substantial part of the cost is the cost of hiding the sprinkler piping and often the heads themselves. This is not required for life safety. In such a case, these panels would very likely be cheaper than tearing into the ceiling to conceal the piping. When I saw such panels installed over a substantial area of a department store, I wondered what the effect of panels’ melting out and falling would have on the fleeing occupants in an emergency. On the one hand, they might serve a purpose by keeping fleeing people out of the fire area. On the other hand, they might panic them into thinking the building was collapsing.


Captain Mike Kopp and Lieutenant Tim Leidig of the Elmhurst (IL) Fire Department share with us this account of a narrow escape from a truss collapse. A roofer working with a torch on the high pitched roof of a restaurant started a fire. As usual, he ran to his truck for a fire extinguisher, which he directed into a roof vent uselessly.

The fire department was unaware of the roof’s construction (see Figure 2). It could not be a better designed firefighter trap. One of the important features of a trap is to disguise it. The roof was a scissor truss used to provide a cathedral ceiling. The original ceiling was gypsum board nailed to the trusses. A later renovation installed a horizontal drop ceiling with batt insulation on top.

Firefighters entered the restaurant, which had moderate smoke. No fire could be seen. Fortunately, the chief of this small department arrived shortly after the first unit and assumed the outside command position. The chief noticed the smoke and fire coming from the roof and ordered an evacuation. As firefighters exited, the roof collapsed.

Departments should train the “two-out” personnel to be observant of hazardous designs and communicate them to the “two-in” personnel. They should be proactive in their safety function, not just passively waiting for something to happen.

The department has now conducted a thorough survey of its fire response area to determine truss locations and has changed its operations to require an outside command, a position from which to see hazards unknown to the interior commander.

A thermal imaging camera should be available to all first-alarm units so they can use it to look for hidden fire overhead.


People working on roofs have started a number of serious fires. Two very costly fires were started by contractors cutting holes for skylights in combustible metal deck roofs, to reduce energy costs. Practically all such groups have cell phones today. Fire departments should identify all contractors, plumbers, and paint removers, especially those who use heat-producing devices, and instruct them in how serious it is to delay alarms; urge them to call 911 immediately.


Wooden I-beams were used as roof rafters in townhouses in Silver Spring, Maryland (see photo on page 80). Firefighters, aware of the fact that trusses make that attic space useless, might think, “The attic space is used. No trusses. No problem.” The wooden I-beams have the same fast burning characteristics as trusses. Do not think the design makes each one a firestop. It is legal to punch holes in the I-beams for air ducts, wiring, and so forth. The industry has put out a video comparing firefighters’ reluctance to accept wooden I-beams with old-time firefighters’ not accepting breathing apparatus. Alice in Wonderland lives on!


Commercial tenants usually seek the maximum open space for flexibility. This means large, open areas, much of which are beyond the reach of the typical high-rise pack’s 13/4-inch hose-in some cases, beyond the reach of 21/2-inch hose. The Wilmington (DE) Fire Department developed a respectable stream with a lightweight portable gun fed by two 13/4-inch lines.11 In some cases, even heavier streams have been found necessary. If you have such large, open areas, now is the time to think about how you would handle the fire. Don’t dismiss the problem if the area is sprinklered. The probability is low, but the sprinkler may be shut off for repairs or alterations at the time the fire occurs. I have recommended that each unit carry one length of 21/2-inch hose, bundled, so that the big line can be made up and transported in a hurry.


The problem is becoming evident in residential buildings. One page 531 of BCFS3, there is a picture of an apartment house with its roof collapsed. The caption notes that the multiple partitions may catch the trusses. That was then. This is now. The bottom chord of such trusses [which carries the tensile (pulling) load] is made up of a number of pieces of wood spliced together end-to-end with metal gusset plates. Failure of any connection might cause total collapse. The counterargument is often “one failure would not cause collapse because the trusses are nailed to the ceiling or floor, which acts to transfer the load to the trusses.” This could be valid if you just pulled the connection apart. However, in a fire, all the trusses will be burning, and other connections close to the point of failure would be “pushed over the edge.” Sometimes, Flinch plate girders12 are used. The loss of wood will permit the steel to buckle.

The availability of long lightweight trusses enables builders to offer wider spaces in residences, both individual and multiple. The conventional tactic of pulling a ceiling, which developed when sawn joists, which limited the fire area, were used, can cause an oxygen-starved fire in the truss void to light up and collapse over a wide area. It is imperative that a thermal imaging camera be used to determine if there is fire in the overhead before operating in the area. Never forget that a fire can be raging overhead without being evident.


Michael Terpak has presented an excellent article on cutting roofs on page 54 of the May 2000 issue of Fire Engineering. With respect to combustible metal deck roofs, he says, “You might want to reconsider working over that area.” I will give you the best reasons to reconsider: Your cut is useless, and being on the roof is very dangerous.

Note that if the ladder rests on the canopy, firefighters above this point will cause the ladder to pivot on the canopy and lift the legs off the ground. Determine whether you can quickly make a large hole in the canopy and then put a ladder up through the hole. An extension ladder might work better through the hole than a 35-foot, single-section ladder. (Photo by Peter Connolly, Fire Department of New York)

In tests at Factory Mutual after the famous General Motors Combustible Metal Deck Roof fire in 1953, a fire in the combustible metal deck roof of a test building 20 feet wide and 100 feet long, with no contents, required a 56-square-foot hole to provide an adequate vent. The fire under the roof is moving faster than a firefighter can run and is weakening the bar joists.

Six Dallas (TX) firefighters plunged into the fire from a metal deck roof that expanded and pushed down the concrete block wall. Fortunately, they were near the edge and were recovered. Had they been in the center, they would have died.13

An unsprinklered Pennsylvania K-Mart had a metal deck roof. Firefighters who were about 30 feet into the store noted heavy, black smoke and fire burning under the roof. (The gases generated from the heated tar adhesive/vapor seal, not the contents.) The roof was collapsing. They evacuated immediately. My knowledge of such roofs began in 1946.14

Do not assume that what you see when looking at one part of a roof is representative of the whole roof. Many sawn joist floors and roofs have been replaced in whole or in part with wooden I-beams. Firefighters should not be on or under burning trusses or I-beams. Since you are unaware of which way the trusses or I-beams run, you cannot be sure you are not on failing trusses even if you are away from the obvious fire area. Collapse is usually total. Once a truss, wooden I-beam, or combustible metal deck roof is burning, all personnel should be removed from the roof. Ventilation is no longer necessary because firefighters should no longer be inside the building. Operating on or under burning trusses is playing Russian roulette with a fully loaded gun. If a building is lost, we didn’t lose it. The owner built it to be lost. We have no obligation to sacrifice lives to indifference or ignorance.


Ribbed awnings, installed on a number of older one-story commercial occupancies to improve their appearance, may do the following:

  • conceal building facades,
  • block ladder placement,
  • destabilize portable ladders (see photo above),
  • render smoke conditions unreadable, and
  • block fire escapes.15


Add the Fire Engineering Web site to your list of favorites. When, from time to time, I refer to an article published some time back, you won’t have to search through dog-eared copies in your rec room. Just go to the Web.

The building is your enemy. Know your enemy.


1. I heard the terrible news while having breakfast in the officers mess at Naval District Headquarters in Balboa, Canal Zone. I had gone to Boston many times to roll with Rescue 3 and was aware of the Cocoanut Grove. The other officers, horrified, peppered me with questions relative to how such a tragedy could happen. In 1943, when I went to Boston to become engaged to Ensign Maurine Romkey, I took the occasion to talk with members of the Boston Fire Department and the National Fire Protection Association about the fire. No mention was made of combustible ceiling tile.

2. Jim McElroy set forth the hazards of combustible tile in NFPA (National Fire Protection Association) Quarterly, July 1949. In the same publication (Oct. 1949), Bob Moulton wrote an article summarizing the hazards, according to P. Bugbee, Men Against Fire (NFPA, 1971), 120.

3. The Atomic Energy Commission chief engineer did not approve of swearing. Offenders had to contribute to a Christmas party fund. I taught the staff to use the name of a prominent manufacturer of cane board as a swear word. The engineers were thus brainwashed when there was no code or rule to quote to eliminate the dangerous material.

4. The Department of Commerce adopted such tests as part of its industry standardization program. The department did not review or approve the tests.

5. Juillerat, E., “The Hartford Hospital Fire,” NFPA Quarterly, Jan. 1962. Carrol Shaw, a prominent and effective NFPA member and state fire marshal of Connecticut, took a serious personal interest in the fire, in which his niece had died.

6. “Report of the MGM Grand Hotel Fire,” Clark County Fire Department, Las Vegas, NV, 1981, v-16.

7. I have been told that the Southern Code can be read to prohibit this practice. It should be forbidden in so many words.

8. Isner, M. and Cahanin, G., “Fire Investigation Report, Elderly House Fire,” Johnson City, Tenn., Dec. 24, 1989.

9. Gelbhaus, R., “Recalling Boulder,” Firehouse, Jan. 1988, 31.

10. News in Brief, Fire Engineering, March 1999, 42.

11. See Building Construction for the Fire Service (BCFS3), Third Edition, 471-472.

12. See BCFS3, 118.

13. A photograph of the firefighters standing on the roof and one taken moments later are shown in Preplanning Building Hazards, Fire Engineering, Feb. 1998.

14. The subject is fully covered in BCFS3, 302-308.

15. Ushko, Michael, WNYF, Fire Department of New York, 1999, Issue 2.

FRANCIS L. BRANNIGAN, SFPE (Fellow), recipient of Fire Engineering’s first Lifetime Achievement Award, has devoted more than half of his 57-year career to the safety of firefighters in building fires. He is well known for his lectures and videotapes and as the author of Building Construction for the Fire Service, Third Edition, published by the National Fire Protection Association. Brannigan is an editorial advisory board member of Fire Engineering.

Wooden I-beams used as roof rafters are as dangerous as lightweight trusses. Work off your aerial tower. Venting the roof increases the fire and destroys the support system. It is not safe to be under an unsafe roof. (Photo by author.)



  • News, dispatches, and updates from the editorial staff of Fire Engineering. Contact us .

  • Francis L. Brannigan , SFPE (Fellow) , was a expert on building construction who devoted more than half of his 63-year career to the safety of firefighters in building fires. He was well known as the author of  Building Construction for the Fire Service, Third Edition  (National Fire Protection Association, 1992), and for his lectures and videotapes. Brannigan was an editorial advisory board member of  Fire Engineering . He was well known as the author of  Building Construction for the Fire Service, Third Edition  (National Fire Protection Association, 1992), and for his lectures and videotapes. Brannigan was an editorial advisory board member of  Fire Engineering and taught at the Fire Department Instructors Conference . He passed away in 2006 .

No posts to display