which are all thoroughly decomposed, are piled up in a dry and ventilated place, with a slight covering of fresh earth to keep down any odor that might arise. After a sufficient interval these heaps are ready for further use, there being no trace, in any portion, of foreign matter, nor any appearance or odor differing from that of an unused fresh mixture of earth and ashes. In this way the material has been used over and over again, at least ten times, and there is no indication to the senses of any change in its condition.

"A sample of this material has recently been analyzed by Prof. Atwater, at the Connecticut Agricultural Station, at Middletown. The analysis shows that it contains no more organic matter than Prof. Voelcker found in fresh earth prepared for use in the closet say about two hundred pounds - nearly all of which organic matter is undoubtedly contained when first made ready for use. In my case there was an addition at a moderate calculation of at least 800 pounds of solid dry matter during the six years' use by an average of four adult persons. Prof. Voelcker's analysis showed that the unused earth contained about twelve pounds of nitrogen. Prof. Atwater's analysis showed that my two tons contained only about eleven pounds of nitrogen. By calculation the 800 pounds of solid dry matter added in the use of my material contained 230 pounds of nitrogen.

"Doubtless the constitution of Prof. Voelcker's sample was somewhat different from the original constitution of my own; but, practically, except perhaps for the addition of a trifling amount of residual carbon remaining after the decomposition, they were about the same, and after being used ten times over, the whole of the 8C0 pounds of organic matter added, including 230 pounds of nitrogen, seem to have entirely disappeared.

"It becomes interesting and important to know what has become of this added matter. That it was absorbed into the particles of the earth is a matter of course, and the result proves that after such absorption, it was subjected to such a chemical action of the concentrated oxygen always existing in porous dry material as led to its entire destruction. Porous substances condense gases — air, oxygen, etc., in proportion to the extent of their interior surface. The well known disinfecting action of charcoal- the surface of the interior particles of which equal from fifty to one hundred square feet to each cubic inch of material, and all of which surface

-

is active in condensing oxygen is due not simply to an absorption of foul smelling odors, but to an actual destruction of them by slow combustion, so that the same mass of charcoal, if kept dry and porous, will continue almost indefinitely its undiminished disinfecting action.

"The earth used in the closet is a porous material, sufficiently dry for the free admission of air or of oxygen. The foulest materials when covered with dry earth, at once lose their odor, and are in time as effectually destroyed by combustion (oxygen) as though they had been burned in a furnace. The process is more slow but none the less sure; and it is clear that in the case of my dirt heap the foul matters added have thus been destroyed. The practical bearings of this fact are of the utmost importance. Earth is not to be regarded as a vehicle for the inoffensive removal beyond the limits of the town of what has hitherto been its most troublesome product, but as a medium for bringing together the offensive ingredients of this product, and the world's great scavenger, oxygen. My experiment seems to demonstrate the fact that there is no occasion to carry away the product from the place where it has been produced, as, after a reasonable time, it has ceased to exist, and there remains only a mass of earth which is in all respects as effective as any fresh supply that could be substituted.

"The quantity necessary to provide can be determined only by extended experiment; my experiment proves that the amount needed does not exceed one thousand pounds for each member of the household, and that this amount once provided will remain permanently effective to accomplish its purpose.

"With a suitable public supply of water for the purpose, and with a suitable means of disposal, nothing can be better, and nothing is more easily kept in good condition than well regulated and properly ventilated water-closets. Where these are available, with enough water for their flushing, their use is to be recommended. When there is not sufficient water, there a well regulated system of earth-closets seems to be imperatively demanded. By one process or the other we must prevent the fouling of the lower soil, and the consequent tainting of wells and springs, and the ground under houses and adjoining their cellars. With a system of sub-irrigation pipes which deliver foul matters into earth that is subject to the active operation of oxydizing influences, we need fear no con

taminating of the deep and unaerated soil. It would be better, however, where this system is used, for the disposal of the out-flow of soil pipes, to avoid the use of wells. As a general rule, it is safer not to use for drinking purposes the water of any well near a house or a stable-practically, it is better not to use wells at all as a source of water for domestic supply. Filtered cistern water is greatly to be preferred."

THE VALUE OF VITAL STATISTICS.

BY PROF. JOHN E. DAVIES, M. D.,

University of Wisconsin.

In most of the questions of life, even the most momentous, our actions are based upon mere probability. Representing absolute certainty by one, in nine-tenths of our decisions concerning common affairs, we are doubtless determined to act by probabilities whose numerical value does not much exceed one-half. While, in many cases, where great prizes are at stake, or the consequence of failure are very disastrous, we do not hesitate to take up with one chance among a thousand, or a million; and we sometimes feel compelled to act, even if blindly, under the influence of mere hope or fear, being, meanwhile, utterly ignorant as to whether we have a single chance for us, among the infinitude of possible chances for and against us.

It is, nevertheless, always an advantage to know beforehand, if we can, the numerical value of a risk; even, where from the nature of the case, we are unable to change, in any manner, the circumstances that govern this numerical value. We can thus act intelligently, even if we choose to act unwisely, and we are able to fix responsibility where it belongs. We trace more readily the connection between cause and effect, and are better able to judge whether supposed causes are such in reality or not.

The numerical measure of the probability of an occurrence is the ratio which the number of chances favorable to it, bears to the total number of chances favorable and unfavorable. When all the chances are favorable, this ratio, of course, becomes one.

In many cases this ratio is fixed and determinate in the nature of the risk, and we may happen to know its value beforehand; or we may be able to calculate it beforehand by more or less simple processes

in arithmetic or algebra. In such cases the probability is called an a priori one. For example, we know beforehand, that in tossing an unloaded coin, the chances in favor of the turning up of a head at a given toss, are as one to two; or that, in the throw of a die, the chances for an ace at a single throw are as one to six. So also the chances in various games, are fixed in their value by the total number of cards in the pack, the number of cards of each kind, the number of persons playing, and the adopted rules of the game. Thus in a game of whist we may readily ascertain that the a priori probability that the dealer has precisely three trumps besides the trump card, is 4,662 divided by 15,875, or that the chances for it are somewhat less than 1 to 3; that the chances for "honours" are as 650 to 1,666; that there are 13 chances out of 8,192, to get six "by tricks," or 1,716 chances out of 8,192 to get one by the same, etc.

But by far the most important cases among probabilities, are those in which we do not thus know the antecedent probability of the events concerning which we inquire. We are then compelled to resort to experiment, or to the observation and record of a large number of cases under very variable conditions, in order to infer the probabilities in favor of future events, from the observance of those events that have passed. A probability, so ascertained, is called an a posteriori one. For example, in the case of the coin, if a very large number of throws brought down a preponderance of either heads or tails, we should be warranted in inferring the unsymmetrical structure of the coin, and hence, a probability greater than one-half in favor of one side or the other at a single toss; because in all cases where a symmetrical coin has been tossed a very large number of times, and by different persons, the number of heads and tails are approximately equal, and tend nearer and nearer to equality as the number of trials increases.* So, in estimating the probable number of deaths from a given disease, likely to occur during any specified time in a stationary community, we should have to be guided, however imperfectly, by statistics or tables of

An Interesting table of the results of a great many series of trials as to the occurrence o heads or tails in the throw of a coin will be found in De Morgan's Formal Logic, p. 185. In those trials there were 2,044 tails to 2,048 heads- Buffon is, also, there cited as having made a similar trial, with the result of 1.992 tails to 2,048 heads.

In Jevon's Principles of Science will be found a record of the heads and tails occurring in different sets of trials, each set having an increasing number of tosses. In these tables are seen very clearly the tendency towards an equality of heads and tails as the number of tosses was increased.