Search

The possible risks of infection associated with the practice of licking envelopes, stamps and labels were investigated. Although pathogenic bacteria and viruses were not isolated from sample envelopes obtained from various sources, the gums used in manufacture were found to exert a protective effect against death from desiccation on the bacteria and viruses which had been introduced into them.Staph. aureus and Salm. paratyphi Bremained viable for several months in dried films of two out of four gums tested. An echovirus could be recovered from similar films for up to 30 days, and an adenovirus for up to 10 days.Bacterial multiplication occurred in the gum used for the manufacture of postage stamps. The comparison is drawn between the possible consequences of minimal contamination of this gum and of cans of corned beef in the light of recent studies.

During the past year a great deal of attention has been given to the question of contamination of working clothes by mineral oils, and some firms have introduced schemes aimed at reducing the skin hazard from this cause. The basis of all the arrangements known to us is to ensure that all engineering overalls are laundered weekly. Some firms supply each of the workers with two pairs of overalls and have arrangements with laundry companies whereby soiled overalls are collected for laundering once weekly and freshly laundered overalls are delivered. Other firms do not supply the overalls but make them available to the workers at low cost. Some firms bear the cost of the launder ing, in others this cost is borne by the worker, but no doubt where arrangements have been made for regular laundering of bulk quantities the cost of this service to the individual is less than it other wise would be. In yet other firms a small sum of money is included in the weekly pay packet to defray the cost of laundering overalls, the worker making his own arrangements for this. Observation of the state of many of the overalls in one factory where the last mentioned arrangements operate led to doubts about the application of the money supplied to the workers for laundering over alls. A suit of overalls was obtained from one worker on a Wednesday. We accepted his statement that his overalls are laundered weekly and that the suit in question had been brought into use 'clean' on the Monday. The 16 sq ft (1-48 m2) of material which comprised the legs and waist pieces were submitted to solvent extraction and yielded over 3 oz (0-09 kg) of a viscous dark brown mineral oil which contained fine silica particles and minute metal turnings from which it was freed by filtration. This oil was saturated aliphatic in nature and had a strong ultraviolet fluorescence; there can be no doubt that it originated from lubricants and cooling oils used in the metal working process. This finding posed several questions : 1. Where laundering of overalls is left to the worker is it likely that this will be done regularly and efficiently ? 2. Is once weekly laundering sufficient or should the interval be shorter ? 3. How efficient is wet laundering of overalls in removing mineral oil contamination ? To resolve these questions 12 pieces of a cotton fabric known as blue drill, each 1 ft square (009 m2), were obtained from a local overall manufacturer and washed in anionic detergent three times and well rinsed in order to free the material from both synthetic and farinaceous sizes and dressings. They were then air dried down to an approximate normal moisture regain of 7% and thereafter six were inoculated with a light grade of engineering lubricant, corresponding approximately to S.A.E. 10 and six were inoculated with a heavy grade corre sponding to S.A.E. 30. Application took the form of total immersion in a 1 to 6 vol/vol solution of the lubricant in petroleum ether. Air drying was allowed to proceed together with drip drying and on an average it was found that 6-4 g (range 6-19 to 6-72) of the S.A.E. 10 oil and 6-7 g (range 6-2 to 71) of the S.A.E. 30 oil were absorbed by the 1-sq ft samples of the blue drill material. After approximately 24 hours air drying, when the test pieces no longer smelled of petroleum ether, they were wet laundered in hot soft water using a well-known brand of anionic detergent which contained soda-ash, metasilicate and carboxymethyl cellulose as manufacturer's additives. This detergent formula is similar to those recommended by the British Launderers' Research Association. After rinsing and normal drying the test pieces were exhaustively extracted and it was found that the S.A.E. 10 test pieces still contained 3-3 g (range 2-9 to 3-59) of oily contaminant and the S.A.E. 30 pieces 3-9 g (range 3-49 to 4-31)

A problem often encountered in large-scale epidemiological studies is the assessment of a per son's relative adiposity from very limited data such as measurements of his height and weight only. There have been several papers in recent years, those of Billewicz, Kemsley and Thomson (1962), Kemsley, Billewicz and Thomson (1962), Khosla and Lowe (1967), Evans and Prior (1969), and Florey (1970), which have examined the properties of height-and-weight indices without agreeing on a single ideal index. The present paper develops a unified theory and gives some examples of its use.

This letter was written in criticism of the Correa-Beasly article entitled "Mathematical Models for Decision-Making in Population and Family Planning" published in the January 1971 issue of the Journal. The author believes that while that article claimed to have used linear and quadratic programming in connection with the acceptance of various contraceptives they did not specify an objective function which is the first principle of linear programming. They did not mathematically formulate their stated objective of minimizing the number of pregnancies through the use of direct contraceptive techniques. Another shortcoming of the paper is that its authors failed to recognize the possibility of substitution of one contraceptive for another. The author also finds fault with some mathematical formulations and their application. Finally he claims that he has recent articles in the same area which more clearly approach the problem although in a different way.

THE PROCEDURE discussed in this paper pertains to the situation in which disease frequency or disease severity can be expressed by several variables, or by several rank categories of the same variable. For example, improvement of a disease under treatment can be measured by change in x-ray, hematological measurements, biochemical measurements, and the like, and the clinician wishes to know which of the several choices he should use to express treatment response; or, an epidemiologist studies a disease that occurs in several degrees of severity, and he wishes to weight the several categories in such a way that differences between population groups attain maximum values. The procedure most often selected is that of choice; for example, the clinician will choose one of his measurements to demonstrate the effect of treatment; or, the epidemiologist will either choose one of his severity categories, for example, mortality, or, in case of low mortality, assign an arbitrary weight to each of the severity categories. In doing so, investigators are often aware that a considerable part of their collected data could be lost, and will feel a need for evaluating the magnitude of the loss. Information is measurable by simple statistical methods. If a variable, Y, and a variable, X, have a correlation coefficient, _K, then r2,. is that fraction of the variation in Y that can be "explained by" variable X. The multiple correlation coefficient, R27.123 is a measure of the amount of information that several variables, X1, X2 and X3 say, contribute to the variation in Y. Several variables may be interdependent, and they often are in clinical and epidemiological studies. The significance of the independent contribution of one variable, X3 to Y, is then judged by the additive information that this X-variable gives above that contributed by the other variables (R2,.123-R2Y.12) In using appropriate scoring for several dependent variables, a factor is assigned to each variable and the weighted sum is used as a single dependent variable, for example: Let cl Y1+c2 Y2 + . ..=u; then, the weighting coefficients, c, are determined in such a way that the multiple correlation coefficient of u on the several X-variables is maximalized. The method has developed from several sources and is in one form discussed by Fisher in "Statistical Methods for Research Workers."" A theory of sig