FLAMMABILITY OF PROPELLANTS 499 (2) In this country a lady was using a hair spray packed in a squeeze bottle. It is believed that the bottle was not of a material recommended for this purpose. It split, spilling lacquer down her dress which was ignited by an electric fire. (3) A squeeze bottle was left in front of an electric fire in a hairdresser's shop. It exploded and flashed. Some damage to the ceiling and walls of the shop occurred. The latter two of these incidents are due to misuse of conventional products generally regarded as being safe under normal circumstances. The first savours strongly of a similar history. No doubt many similar unfortunate incidents have occurred with other common household articles, and while appropriate precautions should be taken, and warnings given, I can see no reason for treating aerosols as a special case per se. I do think, however, that appropriate tests should be applied to a pressure pack to assess the risk associated with it, and that these tests should be appropriate to the product. DISCUSSION MR. G. F. PHILLIPS: In connection with your flame extension experi- ments, using mixtures of 100/120 ø petroleum distillate with butane and water, (1) would you care to comment on the remarkable increase in flamma- bility (0 to 50 cm flame extension) on progressively discharging a spray containing 70% flammable material? (2) Have you any data regarding the apparent fractionation of these formulations? (3) What significance do you attach to the crossing of the curves for full, half empty and 90% exhausted packs originally containing over 70% flammable components,? The difference in flame extension at 80% is small but the three curves appear reasonably contiguous. (4) Do you consider it legitimate to classify hydrocarbon gases together with all solvents present possessing flammable yapours? Their behaviour in the emitted spray and interaction with other, non-flammable, ingredients may be quite distinct. MR. P. DYSON: It would appear from Figure 2 that a full container with 70% butane, 30% water and an emulsifier, with the type of valve used on this type of product (vapour phase tap, etc.) gave no flame projection. Is this in fact so? We have obtained a flame projection of 35-45 cm with such a mixture with a conventional valve, without a vapour phase tap. It is note- worthy too, that 80% butane/20% water in Figure 2 gives a flame pro- jection approaching 55 cm. In my opinion the caption to Figure 2 is most misleading. The graphs do not compare the flame extensions of butane and Arcton-propelled pressure

500 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS packs of simi!ar types. Rather, they compare flame extension of Arcton propelled packs containing a highly flammable matehal with those of butane- propelled packs containing water. THE LF. CTURER: On a point of detail in 3gr. Phillips' question: I have not used 100/120 petrol with butane and water. This was used with Arcton- 11/12 mixture, and the butane/water mixture formed a separate series of tests. However, his first point is similar to that of 3gr. Dyson and serves to demonstrate the importance of correct formu!ation. It also demonstrates the importance of correct testing. I believe that with the mixture containing 70% butane and 30% water very rapid settling of the water in the emulsion has occurred in spite of shaking the dispenser, and most of the water was used up early in the life of the dispenser. This point should have been made explicitly in the text. 3gr. Dyson's emulsions were, no doubt, more stable than those described. I believe that this also covers 3gr. Phillips' second question. I do not think that any great significance should be attached to the cross- ing of the curves relating to the butane packs. It is difficult to measure the flame length with accuracy, and I suspect that all are in fact converging to the same asymptote. I cannot agree with 3gr. Dyson's comment on the caption of Figure 2. This shows flame extensions found with two series of packs containing the same ratios of flammable and non-flammable materials. In one instance, they are Arcton and petroleum ether in the other they are water and butane. It is quite true that some of these mixtures are unlikely to be of any great commercial interest, and this applied equally to both types insofar as I am aware. If taken to the extreme one might say "because a pack contains Arcton it must be safe" and "because a pack contains butane it must be dangerous". Either of these bald statements is untenable, one no more so than the other. The fallacy is, I think, in the tacit assumpt/on that there is some magic difference between propellants and solvents, or other ingredients. I strongly suggest that the only difference between the two classes lies in their vapour pressure, and that even here the dividing line is not well defined. This takes us to 3gr. Phillips' final question. I consider it not only legitimate, but essential, to consider hydrocarbon gases together with all the other ingredients present, not only those with flammable yapours, and that the behaviour of the complete pack should be considered as a whole. MR. A. HERZI•A: Would someone care to comment on the remarks made in the paper that "leakage from pressurized packs presents no greater hazard than from other common household articles"? Gr. F. PHILLIPS: The lecturer quotes four (presumably U.K.) in-