Page images

of 710; the Rockaway watershed, supplying the city of Jersey City, a population of 157 per square mile; the Delaware River watershed, supplying the city of Chester, 173 per square mile; the Merrimac watershed, supplying the city of Lawrence, a population of 69 per square mile.

So far as this particular item of population goes, it can readily be seen that for this case, that of Birmingham, the population on the watershed is among the lowest of those recorded, and that no great danger of pollution is to be expected on the score of excessive population.

In addition to the question of the number of people residing on the watershed, there is always the possibility of accidental pollution. This potential danger particularly leads to the needs for filtration and sterilization of water supply. The likelihood of such pollution is best measured by the presence of the B. coli in the water.

The records of the occurrence of B. coli in one cubic centimetre of water are shown in Table IV. The records are given for the four years from 1910 to 1913. The raw water shows roughly eighty per cent. positive occurrence of B. coli in one cubic centimetre of water by the presumptive test. This water is filtered before being supplied to the consumers. As compared with other water under similar conditions, the water is much the same as that of the Passaic River supplying the Little Falls plant of the New Jersey Water Company. It is somewhat better than the water of the Hudson River reaching the Albany filters, and somewhat better than the water of Brandywine Creek in the Wilmington filters, and somewhat better than the water supplied to London for filtration.

After filtration there was from o to 9.7 per cent. of positive tests showing B. coli in one cubic centimetre. This, as compared with other waters, under similar conditions, shows very well. It is somewhat better than some of the waters, somewhat worse than others.

On the basis of typhoid fever death-rates, which are often taken as an indicator of the quality of the water supply, it is to be noted that the city of Birmingham, based on the 1910 population, showed in 1910 a death-rate of forty-eight per 100,000 population; in 1911, forty-seven; in 1912, forty-two; in 1913, fortyfive. These rates are, of course, relatively high, but compared

with some twenty-nine southern cities of the larger size, having an aggregate population of something like two million people, the average of these twenty-nine cities showed an annual deathrate of fifty-seven per 100,000 population. In 1911, out of these twenty-nine southern cities, sixteen had a higher typhoid rate than Birmingham.

Birmingham conditions, as compared with other cities, have been illustrated in the data already shown. (See Fig. 5.) The explanation of the typhoid fever rate at Birmingham is not to be found in the water supply, but in the conditions of unprotected privies and other similar sources of pollution, whose infected contents are spread mostly by flies and mosquitoes.

A rather interesting feature of the water supply at Birmingham is the fact that the water from the Five Mile Creek watershed is sterilized after being filtered and shows no B. coli as delivered to the consumer. Water from the other watershed, the Cahaba, is not sterilized, but it is filtered, and shows an appreciable number of B. coli. Though the B. coli prevalence shown in one case is materially higher than in the other case, the typhoid fever of the two districts is roughly the same.

Certain parts of Birmingham are well sewered, but in these high-grade residential districts the prevalence of typhoid fever during the fly period was about as great as in the unsewered districts.

During the winter, or non-fly period of the year, the typhoid death-rate was low in all parts of the city.

It is logical to conclude that the occurrence of a moderately limited number of B. coli in any water means nothing as to the quality of the water, as opposed to a water which will show no B. coli whatever. It is only when the coli content becomes markedly high that it can be deduced that the water is, in consequence, unreasonably polluted.

As at Jacksonville, it was the fly and mosquito that transmitted typhoid at Birmingham.

The Density of Oxygen. A. F. O. GERMANN. (Journal of Physical Chemistry, vol. xix, No. 6.)—A brief historical survey of the more important determinations of oxygen is made, followed by a detailed account of the modern methods adopted in the Geneva Laboratory in the determination of gas densities, with particular reference to the purification of the gas used. An improvement in the globe method, as it has been developed in Geneva, is described, consisting of the simultaneous use of a number of barometers equal to the number of globes employed. The density of oxygen was revised, operating on gas furnished by the decomposition of potassium permanganate and purified by fractional distillation. Fifteen determinations led to the value of a “normal litre” under a pressure of 760 mm. of mercury at the temperature of melting ice at sea level in latitude 45° north: Ln = 1.42906. This value, taking into consideration the results of Morley and Rayleigh, leads to the value Ln = 1.42905, which represents, at the present time, the most probable value for the weight of the normal litre of oxygen.

The Purchase of Coal on a B. T. U. Basis. From “ Coal for the Navy,” J. O. RICHARDSON. (Journal American Society of Naval Engineers, vol. xvii, No. 2.)—The Navy Department has frequently been urged to purchase coal under strict specifications that provide that the purchase price shall be a certain figure for coal that shows a heating value of a certain number of British thermal units per pound; that coal which falls below or exceeds in calorific value this number shall be purchased at a price less or greater than the base price by an amount depending upon the number of B. T. U.'s below or above the standard. Theoretically this is an ideal way of purchasing coal, but practically it is very unsatisfactory, because it frequently happens that coal which in a laboratory test would be rated very high on a B. T. U. basis gives very poor results when burned in the furnace of marine boilers, and conversely.

In fact, one of the very best steaming coals in the market, and one which is largely used by sea-going steamers, would be barred from competition because of its lower B. T. U.'s than other firstclass coals, and would therefore be forced to accept a lower price than such coals or abandon the navy trade, which it probably would. This coal, George's Creek, has long been recognized by marine engineers as one of the most satisfactory in the market. This method of purchasing coal would no doubt cause many disputes between the suppliers of coal and the Navy Department, and in the end would be unsatisfactory, because the navy is not buying so many heat units, but is buying high-grade coal that by actual use has been shown to be the best coal obtainable for naval boilers, and the suitability of the coal for the duty required should be the controlling factor in the purchase of coal for use on board naval vessels.



Department of Chemistry, Swarthmore College.

Member of the Institute

INTRODUCTION. ALTHOUGH the application of the laws of mass action to a very large series of reactions between ions and undissociated molecules has received excellent verification, and Ostwald's dilution law has been proved to hold for a large number of binary electrolytes, investigation has shown that it apparently fails in the case of the so-called strong electrolytes. Further, with many weak electrolytes, such as mesaconic acid," a-chlorobutyric acid, 2 2, 4, 5-trimethyl-benzoic acid, and a-oximino-butyric acid,4 to select a number of examples at random, the dissociation constant, as calculated by means of Ostwald's dilution law, is found to vary with dilution. Whether the degree of dissociation of these electrolytes is derived from the conductivit

^ ), or from

do de the freezing-point lowering, a = ", where do and dt represent the observed and the theoretical (on the assumption of no dissociation) freezing-point lowerings, the values calculated for

-vary with concentration and deviate from the required constant. It is obvious, therefore, that either the dilution law does not correctly express the behavior of these electrolytes, or that the methods employed for the determination of their degree of dis

* Communicated by Professor Creighton.

White, G. F., and H. C. Jones, Amer. Chem. Journ., 44, 159 (1910).
? Lichty, D. M., Ann. Chemie, 319, 369 (1901)
* Pfaff, A., Dis. Heidelberg, (1897).
* Hantzsch, A., and A. Miolati, Zeitschr. physik. Chem., 10, 1 (1892).



sociation gives incorrect values. This apparent break-down of the dilution law has been the subject of much discussion by Jahn, Arrhenius, and others, and opinions still differ widely 5 as to the cause of its inapplicability.

Since the dilution law holds for such a large number of electrolytes, and since its validity has been substantiated by thermochemical methods, solubility measurements, hydrolytic measurements, and measurements of reaction velocity, etc., it is probable that its failure, in certain cases, is due to the fact that the determination of the degree of dissociation by means of conductivity or freezing-point methods does not give correct results. The reasons for this are both physical and chemical. Variable friction of the ions; the existence in solution of interaction between the undissociated molecules and the ions, which counteracts their mutual independence ;8 the formation of inner complexes, and the hydration of the ions 8 are all factors which tend to vitiate the values obtained from conductivity or freezing-point measurements for the degree of dissociation.

In view of the failure of conductivity or freezing-point measurements to give, in certain cases, correct results for the degree of dissociation, it is desirable to have recourse to other methods. Jahn ' has attempted to determine the degree of dissociation by the measurement of concentration chains, but the method requires great precision and is hardly sufficiently sensitive where ordinary accuracy is required. Rothmund and Drucker 10 have determined the degree of dissociation of picric acid in aqueous solutions of different concentrations by measuring the distribution of the acid between water and benzene, and from the values obtained have shown that the value of the dissociation constant of picric acid varies slightly, but irregularly, with the concentration. On the other hand, the value for the dissociation constant of this acid is found to vary considerably with concentration, when values

Jahrb. d. Elektrochemie, 8, 102 (1902), and A. A. Noyes, Technology Quarterly, 17, No. 4 (Dec., 1904).

Jahn, H., Zeitschr. physik. Chem., 33, 545 (1900); 35, I (1900); 37, 490 (1901); 41, 257 (1902); Nernst, W., Ibid.. 38, 487 (1901).

'Bredig, G., Zeitschr. physik. Chem., 13, 262 (1894); Noyes, A., Ibid., 36, 63 (1901); Steele, Ibid., 40, 722 (1902).

* Blitz, W., Zeitschr. physik. Chem., 40, 217 (1902).

Jahn. H., loc. cit. 1° Rothmund, V., and K. Drucker, Zeitschr. physik. Chem., 46, 827 (1903).

« ՆախորդըՇարունակել »