Burn properties of fabrics and garments worn in India
S. Bawa Bhalla a , S.R. Kale b , D. Mohan c, * a Department of Mechanical Engineering, Cornell Uni6ersity, Ithaca, NY, USA b Mechanical Engineering Department, Indian Institute of Technology Delhi, Hauz Khas, New Delhi 110016, India c Transportation Research and Injury Pre6ention Programme, Indian Institute of Technology Delhi, Centre for Biomedical Engineering, Hauz Khas, New Delhi 110016, India Accepted 10 August 1999 Abstract A full-scale human form dummy was designed for studying the burning of common dress assemblies (i.e. combination of garments) worn in India. The dummy was made in eight parts; each made of a steel shell lled with water so as to replicate properties of skintissue combination. Four thermocouples were xed on the dummy for measuring torso, neck and face temperatures. The dummy was clothed separately in three womens dress assemblies (saree, salwarkhameez and nightgown) and mens dress assemblies (kurtapyjama, shirt pant and lungi), and ignited at the feet by a at ame. The tests showed that loose tting garments burnt easily. Once completely burnt, all dress assemblies result in third degree burns over most of the body. The burning process of synthetics is radically different from cottons or cottonpolyester blends. However, ame duration and temperatures produced on the skin are not radically different, suggesting that on the whole synthetics are no worse than cotton garments. Thick garments, such as, jeans and khadi, do not ignite easily and are inherently safer than similar garments made of light fabric. The studies show that results of standard ammability tests using single fabric strips do not correlate with the burning observed in garments as part of a dress where multilayering is common. Standards/codes for re safety of garments and garment dress assembly combinations need to be evolved to adequately address their re safety. 2000 Elsevier Science Ltd. All rights reserved. Keywords: Dummy; Dress; Dress assembly; Garments; Fabrics; Fire safety; Burning; Burns; Indian www.elsevier.com/locate/aap 1. Introduction The re safety of garments is a subject of consider- able interest and has resulted in standards, interna- tional and country specic, for testing fabrics. In many cases, a test sample, typically 50300 mm, is burnt in a controlled setting. Measurements of heat ux and temperature and ame propagation are used for charac- terising the re safety of the fabric on the basis of certain criteria laid out in the codes. This characterisa- tion procedure has considerable relevance for character- ising the ammability of draperies because the test conditions reasonably mimic actual conditions. These tests, however, do not accurately mimic the conditions of the fabric when it is made into a garment and worn as a combination of garments, i.e. dress assembly. The present study was conducted to explore these aspects, particularly in the Indian environment where an over- whelming number of burn injuries occur at home. Sawhney (1989) analysed 339 patients with burns ad- mitted to a teaching hospital and concluded that 89% of the cases occurred at homes with oor level cooking using a kerosene stove. Two-thirds of these cases re- sulted from mishandling of the stove while wearing loose garments. More recently, analysis by Subrah- manyam (1996) showed that 92% of the burns occurred at home. Similar gures have been reported in other studies (Gupta et al., 1993; Jayaraman et al., 1993; Kumar et al., 1994) in urban areas. Even in a predominantly industrial area, 85% of the burns oc- curred at home where kerosene stove and open ames were common causative agents (Sarma and Sarma, 1994). * Corresponding author. Tel.: +91-11-659-1147; fax: +91-11-685- 1169. E-mail address: dmohan@cbme.iitd.ernet.in (D. Mohan) 0001-4575/00/$ - see front matter 2000 Elsevier Science Ltd. All rights reserved. PII: S0001- 4575( 99) 00112- 8 S. Bawa Bhalla et al. / Accident Analysis and Pre6ention 32 (2000) 407420 408 2. Experimental facility The experimental facility consisted of an instru- mented dummy over which the dress assembly was tted. The re was initiated at the feet with a at ame. Still and video pictures of the burning process were taken besides the temperature history at four locations on the dummy. Details of the dummy and experimental procedures are given below. 2.1. The dummy Conventional dummy designs consist of a multilay- ered shell of bres, usually carbon bres and kevlar, joined together with epoxy (Crown and Dale, 1993; Eaton, 1995). Such a design does not accurately repre- sent the properties of the skinesh combination. Its use is at best limited to short testing periods, less than 1 min, such as those encountered during exposure to ash res or a short run through a re. In prolonged exposures, lasting several minutes, heat transfer to the skin and tissue is important as it affects the ame itself and determines the extent of injury. Fires from burning dresses last several minutes, which necessitated a dummy, whose heat transfer properties matched those of human skin and esh. A review of literature for obtaining thermal proper- ties of the skin and tissue was undertaken, on the basis of which, data given by Bowman et al. (1975) and Duck (1990) was adopted. The data indicate that thermal conductivity and diffusivity of skin and tissue are al- most the same as the corresponding properties of water. This observation can also be traced to the fact that water is the predominant constituent in skin and tissue. It was, therefore, decided to make a composite dummy with a thin outer shell containing water. Several candi- date materials were examined for thermal properties, strength, ease of fabrication, re resistance, cost and availability, after which galvanised mild steel sheet of thickness 0.56 mm was nally selected. For ease of fabrication and dressing, the human form was realised by assembling eight pieces, viz, two each for each arm, one for each leg, torso and head (and neck). The shape of each piece was an approximation of the correspond- ing body portion, such as, cylinder, frustum of a cone or elliptical. The individual parts were assembled with bolts and the bottom of the torso was placed on a steel rod anchored vertically in the oor. The assembled dummy was enclosed in a wire mesh cage with exhaust hood and fan at the top. A photograph of the entire assembly is shown in Fig. 1a. Each piece was lled with deaerated water so as to prevent bubble formation on the inside surfaces. Experiments conducted by Bandhu and Singh (1994) on a single piece showed that over the range of room temperatures, which includes body tem- perature, water temperature did not inuence the fabric combustion. Hence, water in the dummy was at room temperature during the tests. On the basis of a heat transfer analysis it was decided that internal inserts in the parts would not be necessary because natural con- vection currents were insignicant. Fig. 1. (a) Experimental facility the dummy. (b) Experimental facility instrumentation. S. Bawa Bhalla et al. / Accident Analysis and Pre6ention 32 (2000) 407420 409 Fig. 2. Burn sequence of cotton saree dress assembly. Fig. 3. History during burning of a cotton saree dress assembly. S. Bawa Bhalla et al. / Accident Analysis and Pre6ention 32 (2000) 407420 410 Fig. 4. Sequence of polyester saree dress assembly. Fig. 5. Temperature history during burning of a polyester saree dress assembly. S. Bawa Bhalla et al. / Accident Analysis and Pre6ention 32 (2000) 407420 411 Fig. 6. Burn sequence of silk saree dress assembly. Fig. 7. Temperature history during burning of a silk saree dress assembly. S. Bawa Bhalla et al. / Accident Analysis and Pre6ention 32 (2000) 407420 412 Fig. 8. Burn sequence of a cotton salwarkhameez dress assembly. 2.2. Test procedure Each test commenced by selecting the dress (de- scribed in Section 2.3) and cleaning the dummy with carbon tetrachloride. The dummy was then clothed. In order to avoid disturbing the thermocouples, some garments had to be cut and stitched back later. For light up, a U-shaped sheet metal tray, 5050 cm overall was covered with jute fabric and sprayed with kerosene oil. When ignited, it produced a at ame 5070 mm in height at which point it was slid along the oor directly under the dummy. After appearance of a stable ame on the clothed dummy, the tray was re- moved and its ame put out. The burning on the dummy was recorded using still and video cameras and temperatures of the four thermocouples were recorded on the PC. After the combustion was complete, the dummy was cleaned and readied for the next test. This procedure of initiating combustion was used for three main reasons. Firstly, it replicates the situation in an Indian home described in the previous paragraph. Sec- ondly, from a combustion stand point, the ame speed for a burning fabric is greatest when the ame travels vertically upward; representing a worst case scenario for burn injuries. Finally, the at ame provided a well controlled method for ame initiation, something not possible if the ame were to be applied at any other location on the dummy. 2.3. Garments and fabrics 2.3.1. Dress assemblies and garments The experiments focussed on the fabrics, garments and dress assemblies typical to India. Three typical The dummy was instrumented with four thermocou- ples (K-type with breglass insulation) brazed on the head (face), neck, upper torso and lower torso, as shown in Fig. 1b. Prior to brazing, each thermocouple was calibrated up to 180C. These thermocouples were connected to a PC-based data acquisition system via compensating cables. All the thermocouples were sam- pled at 2-s intervals during tests. The dummy was placed upright on a steel rod grouted into the oor. The vertical upright position used here simulates a worst case scenario, in compari- son to lying at or sitting positions, from burn injuries stand point. It is here that the garments are free from the body making them extremely vulnerable to ignition and subsequent combustion. In Indian homes which is where most burn injuries occur, such a position would occur when a person cooking food on the oor stands up and the lower garments are exposed to the open ame of a stove. Fig. 9. Temperature history during burning of a cotton salwarkhameez dress assembly. S. Bawa Bhalla et al. / Accident Analysis and Pre6ention 32 (2000) 407420 413 Fig. 10. Burn sequence of a cotton nightgown dress assembly. Fig. 11. Temperature history during burning of a cotton nightgown dress assembly. S. Bawa Bhalla et al. / Accident Analysis and Pre6ention 32 (2000) 407420 414 Fig. 12. Burn sequence of a thick-khadi kurtapyjama dress assembly. Fig. 13. Temperature history during burning of a thick-khadi kurtapyjama dress assembly. S. Bawa Bhalla et al. / Accident Analysis and Pre6ention 32 (2000) 407420 415 Fig. 14. Burn sequence of cottonpolyester blend shirt pant dress assembly. Fig. 15. Temperature history during burning of a cottonpolyester blend shirt pant dress assembly. S. Bawa Bhalla et al. / Accident Analysis and Pre6ention 32 (2000) 407420 416 womens dress assemblies, viz, saree, salwarkhameez and nightgown, and three typical mens dress assem- blies, viz, kurtapyjama, shirt pant and lungi were tested. The saree dress assembly, Figs. 2a, 4a and 6a, con- sists of ve garments, panty, petticoat, saree, bra and blouse. The saree is a 5 m long and 1.5 m wide cloth which is wrapped twice around the waist, with pleats in the front, and thrown across the chest and left shoulder with the end hanging freely at the back. The petticoat and saree are loose tting whereas the undergarments and blouse are tight tting. The salwarkhameez dress assembly, Fig. 8a, consists of panty, bra, khameez (a knee length dress), salwar (baggy trousers snug at the bottom) and a dupatta (a scarf like cloth covering the chest and thrown across the shoulders). Except for the undergarments, the other garments are loose tting. The nightgown dress assembly, Fig. 10a, consists of nightgown, similar to those worn worldwide, and panty and bra. The mens dress assembly kurtapyjama, Fig. 12a, consists of a kurta (a knee length dress), pyjama (baggy trousers loose at the bottom), vest and brief. The dress is loose tting all over, except for the undergarments. The shirt pant dress assembly is the same as it is the world over; a shirt tucked into a pant, briefs and vest, Fig. 14a. The lungi dress assembly, Fig. 16a, consists of a vest, a brief and 2 m long and 1.5 m. wide cloth tied at the waist and loosely hanging down. 2.3.2. Fabrics A wide variety of fabrics are in use in India, with the cost being an important factor in their use. Besides established large mills where quality records are prop- erly maintained, there are a large number of small and medium mills where quality records are not maintained as well. The latter are lower in cost and, hence, used by a majority of the people. Detailed analyses of the fabrics were, therefore, difcult to obtain and broad classications were followed. All undergarments were Fig. 16. Burn sequence of a cotton lungi dress assembly. S. Bawa Bhalla et al. / Accident Analysis and Pre6ention 32 (2000) 407420 417 Fig. 17. Temperature history during burning of a cotton lungi dress assembly. of cotton. Saree, kurtapyjama and nightgown fabrics were either cotton, polyester or cottonpolyester blend, and silk (saree only). Mens kurta-pyjama were made of thin khadi, thick khadi and cottonpolyester blend. Khadi is a cotton fabric, which is produced on small or medium scale, usually coarse in texture. Shirt pant combination consisted of cotton or cotton-polyester blend shirt, or cotton, cottonpolyester blend or wool T-shirt with either cottonpolyester blend pant or cot- ton jeans. The lungi was made of cotton fabric which was not of khadi variety. 3. Experimental results Multiple runs were performed with each dress assem- bly using different fabric combinations and the results are presented below. Pictures and data, typical of the particular dress assemblyfabric combination are also presented. 3.1. Womens dress assemblies 3.1.1. Saree dress assembly Figures 2ad show the burning sequence of a cotton saree dress assembly and it can be seen that even while the ames exist at the feet, the re has spread along the saree up to and above the face/head (Figs. 2bc). Flame propagation is rapid for loose cloth whereas the tight tting garments and the blouse have burnt par- tially or not at all (Fig. 2d). The temperature history, Fig. 3, shows that after an initial delay, the inuence of ames is felt rst at the head and neck and then at the lower and upper torso. At 140 s, a pair of peaks appears at the lower and upper torso, a result of delayed burning of waist level fabric the panty, petticoat band and saree tucked into the petticoat. Heavy charring was observed which smouldered for a considerable time. The burn sequence of a polyester saree dress assem- bly is shown in Figs. 4ad. After light up, the ames gradually progress upwards, burning the saree fabric below it. Flames do not engulf the head even when torso-level fabric is burning. Chunks of molten burning fabric fall off the dummy to the ground where burning continues (Figs. 4b and d). Figure 5 shows the temper- ature history during this process. Here, the four loca- tions record peak temperatures at almost the same time. The highest temperature occurs at the lower torso and it is substantially greater than the others. Hot gases produced by combustion at the torso ow upward past the face, resulting in simultaneous peaks. An interesting feature is the absence of a peak at the neck and the head due to combustion of chest level fabric. The burn sequence of a silk saree dress assembly is shown in Figs. 6ad; here immediately after ignition, the entire saree is in ames, typical of a ash over. Pieces of burning fabric fall off to the ground where the combustion continues. The temperature history, Fig. 7, shows rapid temperature rise at all locations with the lower torso experiencing elevated temperatures for a much longer period. The secondary peak at the lower torso results from delayed combustion of waist level fabric, which is worn tightly. Test results for cottonpolyester and terylenecotton sarees exhibited similar trends as the cotton saree and, hence, are not presented here. S. Bawa Bhalla et al. / Accident Analysis and Pre6ention 32 (2000) 407420 418 3.1.2. Salwarkhameez dress assembly Figures 8a and b show the burn sequence of a cotton salwarkhameez dress assembly. Hot gases produced by combustion at the feet, balloon the dress which starts burning at several locations. The ames rapidly engulf the head and face, Fig. 8b. The temperature history of Fig. 9 shows rapid ame spread and all locations experience high temperatures for similar time periods. The burning of a polyester salwarkhameez dress assembly exhibited the ballooning observed in the cot- ton khameez but the ame propagation was gradual, much like the polyester saree. Due to variations in localised burning, the temperature histories showed wide variations high temperatures if burning oc- curred in the thermocouple vicinity, moderate tempera- tures otherwise. The cottonpolyester blend salwar khameez too exhibited khameez ballooning and its subsequent combustion was similar to that of the cot- ton salwarkhameez. 3.1.3. Night dress assembly Figures 10ad depict the burning of a cotton night- dress assembly. Immediately after initiating the re to the bottom of the nightgown, it ballooned resulting in rapid ame spread vertically up to the face, along the front in this case. The torn nightgown then burnt slowly all around. The result of this burning is a rapid temperature rise at the lower torso and at the head, seen in Fig. 11. The same trends were observed in the burning of a cotton-polyester blend nightgown. The burning of the polyester nightgown was similar to the saree, gradual progress to the top with pieces of molten burning fabric falling to the ground. 3.2. Mens dresses assembly 3.2.1. Kurtapyjama dress assembly Figures 12ad show the burning sequence of a kurtapyjama dress assembly, the kurta and pyjama being of thick khadi fabric. Qualitatively, this dress assembly has semblance to the salwarkhameez and here also the kurta balloons and ames erupt at the upper portion. The burning is followed by heavy char- ring. The temperature history, Fig. 13, shows a consid- erable delay in re initiation, the result of a snug tting pyjama. The head experiences highest temperatures al- though the duration of elevated temperatures at all locations is comparable. Kurtapyjama sets of thin khadi exhibited similar behaviour, with one difference; the time for ignition was less than that for thick khadi. With both fabrics, vertical ame propagation dominated horizontal ame propagation, which left columns of charred fabric be- hind them. 3.2.2. Shirt pant dress assembly A unique feature of all shirt pant dress assemblies was the difculty in initiating the re. In some cases a re would self extinguish itself after the re tray was removed necessitating re-introduction of the re tray. A sequence of burning of a cottonpolyester blend shirt and pant is shown in Figs. 14ad; the ames gradually progress to the top consuming the pant and shirt in their wake. Differences in the combustion on left leg (Fig. 14b) and right leg (Fig. 14c) are attributed to the difculty in initiating a stable ame simultaneously on both legs, an observation common to this dress irre- spective of fabric type. The temperature history, Fig. 15, shows a peak at the head at about 120 s which is attributed to the burning shirt at chest level. Another peak is observed at the lower torso at about 330 s which is attributed to the burning of waist level fabric of the pant, shirt, vest and brief. The total burning time is quite long, over 7 min. Cotton and wool shirts with cottonpolyester blend pant exhibited similar behaviour. Cotton jeans were very difcult to set on re and, subsequently, the rate of combustion and ame spread were very slow compared to pants of any other fabric. 3.2.3. Lungi dress assembly The burning sequence of a cotton lungi dress assem- bly is shown in Figs. 16ad. The lungi itself burnt very rapidly, the waist level cloth and vest burning compara- tively slowly. The major combustion occurred below the waist and, consequently, temperatures do not attain high values as seen in Fig. 17. 4. Discussion The series of tests described above give insights on the combustion of different dress assemblies and fabrics there in. The observations can be generalised on some important issues and these are presented below. 4.1. Effect of t and multilayering In the dress assemblies investigated, the saree is the loosest tting garment; at the other end of the spectrum is the shirt-pant assembly. The effect of t can be observed on two aspects. First, loose t garments catch re easily and, second, once on re the ame spread is rapid. Both aspects limit the safety of the dress. In all the tests, the dummy was clothed in undergarments that were tight/snug tting. In most cases, these garments did not burn completely and, in some, they did not burn at all. There is a possibility that the tight tting garment could shield the underlying skin from the ame. These types of garments can enhance re safety. In this context it is important to note that results from S. Bawa Bhalla et al. / Accident Analysis and Pre6ention 32 (2000) 407420 419 ammability tests (ISO 6940, 1984; ISO 6941, 1984; IS 11871, 1986) which were obtained for all fabrics, cannot be correlated with the burning characteristics and, hence, re safety of a dress. A fabric used as a light/ snug tting garment is inherently safer than the same fabric used as a loose tting garment. Simultaneously, the limitation of standard tests is evident when multilayered fabrics are encountered. The same fabric or combinations of different fabrics in multilayer arrangements give very different burning characteristics. For example, the cotton petticoat, worn with the saree, burns very differently from a cotton lungi, which is almost the same as the petticoat. With a saree, the petticoat burns after the saree covering it has burnt, thus slowing its burning. The petticoat, without the saree would burn rapidly. On the basis of standard tests, using vertical or inclined strips of fabric, it is not possible to predict the burning of the petticoat. This aspect was observed in the experiments performed by Bandhu and Singh (1994). They rst conducted single strip tests as per standards and then studied the effects of t (loose, medium and tight) by wrapping the fabric once over a vertical cylinder and on a horizontal cylin- der. While results of loose t fabric on the vertical cylinder could be correlated with single strip tests, no such conclusion could be drawn about the tight t cases. Tests on the horizontal cylinder had almost no correlation with standard tests. Similar conclusions were observed for multiple layers of fabrics. Hence, the standard tests, using vertical or inclined strips of fabric are very limited and a comprehensive set of tests cover- ing these aspects need to be evolved to predict the burning/ammability of dress assemblies and garments and not just of the fabrics they are made from. 4.2. Relati6e grading of the re safety of Indian dress assemblies and fabrics The re safety and, hence, injury potential, of the dress assembly-fabric combinations have been analysed on the basis of (i) ignition time, i.e. time required to establish a stable ame, and (ii) total burning time, i.e. time required for complete combustion of the dress. This numerical criterion suppresses the physical charac- teristics of the burning processes. On the basis of these, the re safety of the dress assemblies are discussed below. 4.2.1. Womens dress assembly Amongst sarees, silk sarees were not only highly inammable but as a dress assembly they burnt most vigorously and rapidly, followed by terylenecotton, cotton and polyester in that order. Although the total burning time is similar for terylenecotton, cotton and polyester (the most widely worn sarees) was similar; there were marked differences in the physical character- istics of the ames. Cotton and terylenecotton sarees burn with a large ame with rapid horizontal and vertical spread whereas polyester burns with smaller ames in lumps which either stick to the skin or fall to the ground. These lumps continue burning long after the dress had been consumed. The most inammable salwarkhameez was of terylenecotton, followed by cotton and polyester. The same grading was observed for nightgowns also. In all cases, the burning was similar to the corresponding saree. 4.2.2. Mens dress assembly Amongst kurtapyjamas, thin khadi dress burnt faster than thick khadi; the latter was also difcult to ignite. In comparison to khadi dresses, terylenecotton kurtapyjama was much more ammable. Amongst all dress assemblies, the shirt pant dress was the safest. Further, cotton jeans were safest as they did not catch re easily and even after catching re, they burnt very slowly. The cotton lungi burnt very rapidly. Though the re did not affect the torso, neck or head, injuries to the feet and lower abdomen cannot be ruled out. 4.3. Grading the re safety of fabrics In general, thick garments are less susceptible to burning than thin garments. For all dresses, womens and mens, there is no evidence that synthetics are on the whole more dangerous than cottons. There are, however, trade-offs because the burning characteristics of the two are quite different. Synthetics melt as they burn and break off from the garment; as they fall off, these molten pieces could stick to the skin and cause severe injury. In the tests, it was observed that such pieces occasionally stuck to the dummy. A test was also performed on the leg of a pig (obtained from a local butcher) by wrapping a synthetic fabric on it and setting it on re in vertical position. Interestingly, most of the molten burning fabric that came in contact with the skin did not stick to it. It is possible that these pieces would stick to the skin where it is horizontal. 4.4. Burns injury estimates The temperature histories give magnitudes of ex- pected skin temperatures in case of real burns. Based on the data of Guzzetta and Randolph (1983), thresholds for third degree burns are 54C for 30 s, 60C for 5 s or 70C for 1 s. These criteria were exceeded in most cases suggesting third degree burns if the clothes burn completely. This criteria, however, is based on wet burns, i.e. burns caused by hot water. In general, wet burns are more serious than dry burns because heat penetration is signicantly larger in the former case. For dry burns, the heat ux criteria de- scribed in ASTM D4108 or by Diller (1985) for second- S. Bawa Bhalla et al. / Accident Analysis and Pre6ention 32 (2000) 407420 420 degree burns were used and on this basis, too, the temperature histories suggest third degree burns. 4.5. Burn injury mitigation The experimental data also indicates an important aspect, viz, time available for preventive action. This time can be deduced from the ignition delay times obtained from temperaturetime data. The ignition delay is dened as the time required for a stable ame to appear at the lower torso after the light-up tray has been introduced. All womens dress assemblies and the lungi dress for men have short ignition delays, typically 30 s. For other mens dress assemblies, the delay times were generally more than 50 s. Most dress assemblies required ame from the light-up tray for 35 s; the exceptions being shirt pant and shirt jeans which re- quired ame for as much as up to 1 min. The time required for complete burning (total burn time) of the dress assembly was also obtained from the data. For loose tting garments, these times were less than 2 min; there was considerable variation depending upon the type of dress assembly-fabric combination. All shirt pant combinations exhibited total burn times in excess of 5 min. These data while supporting the earlier conclusion of the effect of t on fabric burning also give an estimate of the time available for preventive action. Strategies for loose t dress res should rst aim at natural retardation of the ames for which ballooning should be suppressed. A quick and easy method for realising this is for the victim to lie at on the ground, preferably face downwards. Lying at on the ground also prevents injuries to the face, head and neck, which are most difcult to treat, and which also leave visible disgura- tions. Writhing in pain, a natural human reaction, exacerbates the re and increases burn injuries. Dous- ing the victim with water, as recommended by rst aid manuals, is likely to be more effective for polyester dresses than cottons as the ame propagation speeds are lower and ash over does not occur. On the whole, res from coarse fabric dresses allow more time for preventive action as compared to ne fabric dresses. The tests clearly point to the use of heavy fabric, particularly cotton, dresses in snug or tight t, for maximum re safety. 5. Conclusions The tests conclusively show that standard ammabil- ity tests where single layered strips of cloth are burnt do not accurately predict the burning process when the same fabric is made into garments and worn as part of an overall dress assembly. Most dress assemblies, on complete combustion will result in third degree burns. There is no evidence that synthetics on the whole are worse in re safety as compared to cotton for same garment dress combination. Burn injuries can be min- imised by wearing tight/snug tting garments made of thick fabrics. Consequently, future action should be directed towards public awareness; development of re retardant cottons and synthetics, and drafting of codes/ standards for re safety of garments and dress assem- blies in addition to those for the fabric. Acknowledgements The nancial support of the WHO is gratefully ac- knowledged. References Bandhu, S., Singh, S.P., 1994. Design and Fabrication of a Dummy for Burn Studies. B. Tech. Thesis, Mech. Eng. Dept., IIT Delhi. Bowman, M.F., Carvalho, E.G., Woods, M., 1975. Theory, measure- ments and application of thermal properties of biomaterials. Ann. Rev. Biophys. Bioeng. 4, 4380. Crown, E.M., Dale, J.D., March 1993. Built for the hot seat. Can. Textile J., 1619. Diller, K.R., 1985. Analysis of skin burns. In: Heat Transfer in Medicine and Biology: Analysis and Applications, vol. 2. New York, pp. 9495. Duck, F.A., 1990. Thermal Properties of Tissue, Physical Properties of Tissue. Academic Press, London. Eaton, P., 1995. BTTG Fire Technology Services, Altrincham, UK (personal communication). Gupta, M., Gupta, O.K., Yaduvanshi, R.K., Upadhyaya, J., Febru- ary 1993. Burn epidemiology: the pink city scene. Burns 19 (1), 4751. Guzzetta, P.C., Randolph, J., March 1983. Burns in children. Pediatr. Rev. 4 (9), 271278. IS 11871, 1986. Methods for Determination of Flammability and Flame Resistance of Textile Fabrics, Bureau of Indian Standards, New Delhi. ISO 6940, 1984. Textile Fabrics-Burning Behaviour-Determination of Ease of Ignition of Vertically Oriented Specimens, International Standards Organisation. ISO 6941, 1984. Textile Fabrics-Burning Behaviour-Measurement of Flame Spread Properties of Vertically Oriented Samples, Interna- tional Standards Organisation. Jayaraman, V., Ramakrishnan, M.K., Davies, M.R., August 1993. Burns in Madras, India: an analysis of 1368 patients in 1 year. Burns 19 (4), 339344. Kumar, P., Sharma, M., Chadha, A., June 1994. Epidemiological determinants of burns in paediatric and adolscent patients from a centre in Western India. Burns 20 (3), 236240. Sarma, B.P., Sarma, N., June 1994. Epidemiology, morbidity, mortal- ity and treatment of burn injuries A study in Peropheral Industrial Hospital. Burns 20 (3), 253255. Sawhney, C.P., 1989. Flame burns involving kerosene pressure stoves in India. Burns 15 (6), 362364. Subrahmanyam, M., September 1996. Epidemiology of burns in a District Hospital in Western India. Burns 22 (6), 439442. .