PLASTIC COATED PUSH BUTTON CONTAINERS 389 locity that the rubber becomes brittle and will break rather than stretch. Polyvinyl chloride exhibits these same characteristics. In the manufacture of glass aerosol containers, attempts are made to stay as close as possible to the size and shape indicated as desirable by the cus- tomer. From an engineering standpoint we are able to suggest modifica- tions in bottle designs which will, in the majority of cases, adapt a non- pressure designed bottle to a glass aerosol (which we consider to be a con- tainer that will withstand a minimum hydrostatic pressure of 150 lbs. per square inch for one minute). At this time bottles are produced on a single section of equipment. When the bottle design has been proved to be suitable for aerosol--from the hydrostatic standpoint--we then proceed with drop-testing to determine if peculiarities in a break pattern can be es- tablished. Such peculiarities--if encountered--are brought to the atten- tion of the customer, who evaluates the container according to his stand- ards and rejects or accepts the bottle. When the bottle goes into production a statistical quality control, as outlined previously, is maintained to insure that bottles are being produced to meet the standards specified. The bottles are then run through our bottle coating line where they 1. Subjected to pressure test before coating. 2. Coated. 3. Decorated. 4. Vented. 5. Packed for shipment. Testing the coated bottles involves the following: 1. Simulated customer pack is drop tested at 100øF. to determine if coating thickness is adequate. 2. Weight of deposited coatings is recorded periodically. 3. Weight distribution of plastic jacket is observed for check on ma- chine. There are several factors which make drop testing of bottles subject to considerable variance in evaluation of data. Primarily, because breakage takes place so rapidly no one can observe the break without the aid of high-speed, stop-motion photography. This automatically produces a large measure of subjective conclusion even among scientific personnel who strive to be objective in the evaluation. It is often helpful to drop a bottle filled with water to establish a refer- ence point for evaluation of aerosol drop-testing. A bottle containing a liquid non-pressurized and non-plastic coated will shatter upon dropping. The distance the glass fragments are thrown de- pends upon the size and shape of the bottle, the total weight of the pack- are:

390 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS age, the height from which it is dropped and the surface upon which it is dropped. The same container, empty, dropped under identical conditions will pro- duce substantially different results. The difference can be explained partially by postulating that the residual inertia of the liquid mass is translated into a force which acts against the walls of the container creating a new fragmentation pattern without in- creasing the possibility of bottle breakage, per se. If we assign a category empirically to the typical break pattern of an empty bottle, we would find that a non-pressurized bottle filled with liquid would fall in a different category. If we take an identical container filled with a pressurized liquid, like soda water, and subject it to the same test under identical conditions, we find another category must be added to the two above. Then taking the same container and using an aerosol fill in it, under iden- tical conditions, we find we must create a fourth category. The fragmen- tation patterns of each of the above are substantially different and statis- tically significant. Safety evaluation of an aerosol package without an adequate industry- approved method of evaluation is practically an impossible task. Until such time as a "tool" is p•oduced which will measure hazards objectively any method now in use by anyone approaches something like attempting to measure a mile with a foot rule. Because there are no accepted, industry- wide reference points it often becomes a management decision to decide on the relative safety of a glass aerosol. These decisions are of necessity made with a minimum of data and practically no reference points. For this reason a CSMA subcommittee, under the chairmanship of Dr. Joseph Lann of the Kinetic Laboratory, E. I. du Pont de Nemours and Company, is attempting to provide a method of objective evaluation of hazards involved in glass aerosols. The proposed test is under evaluation by nine cooperating laboratories to determine reproducibility of results and to consider the statistical approaches for drawing objective conclusions. In brief, the proposed test uses a paper tent into which the bottles are dropped from varying heights onto a 1/4-inch steel plate embedded in con- crete. The mean height from which the bottle must be dropped to produce breakage and fragmentation is determined. Hazard is determined by per- forations in the paper tent. Assuming results of the cooperating laboratories to be sut-'ficiently in agreement, tentative minimum standards will be proposed to the industry. This should give the aerosol packer and merchandiser a yardstick by which he can measure his product in an increasingly competitive market. Maximum safe pressure in glass aerosol formulations depends upon a