This may be partly accounted for by the large percentage of water surface contained in the area in reservoirs, ponds, swamps, etc., but even after making a correction for that the amount is still much below the run-off from other areas for the same period.

It is plain that the Sudbury River is not a good dry-weather stream, and that estimates based on the flow of that stream as given in the reports would at least be conservative and safe for most areas in the vicinity.

This but illustrates the fact which can scarcely be too strongly emphasized in closing, — that most areas are to a certain extent a law unto themselves, and have certain peculiarities of their own, which should be carefully studied and compared in the light of the best data obtainable, if the best and most reliable results are to be secured.

DISCUSSION.

PROF. DWIGHT PORTER. Mr. Hastings has referred to the formula by Mr. Vermeule of New Jersey, by aid of which the evaporation from land surfaces can be computed. I have a friend here from California who tells me that in a case there with which he is familiar they applied that formula to the evaporation from land surfaces, with the result that the evaporation for the year was shown to be much more than the entire rainfall for the year, which seems to be an absurd result, and indicates, I think, the danger of putting too much confidence in a formula of this kind, especially when you apply it to a region distant from that for which it was framed.

MR. HASTINGS. I think Mr. Vermeule in his book expressly states that the formula is applicable primarily to New Jersey, but he thought it might be applicable to many other localities. I doubt, however, whether he would want to apply it to California or Arizona.

MR. RICHARD A. HALE. I have been very much interested in Mr. Hastings' paper, and I think there are a good many important facts in it. He has spoken of the different yield of large and small drainage areas, and I think that is very noticeable in the observations taken on rivers with drainage areas of different sizes. At Lawrence, on the Merrimac, which is one of the rivers

which Mr. Hastings refers to, we have a minimum yield of .3 of a cubic foot per second per square mile during the dry season; and on the Shawsheen and Spicket rivers, which are adjacent, with areas of only about 70 and 84 square miles respectively, the minimum yield is about one-third of that, or about 0.1 cubic foot per second per square mile. We are having some daily observations made on these rivers, and the difference is particularly noticeable on small streams. During the year 1900, which was an exceedingly dry period, the minimum weekly flow of the Merrimac was .3 of a cubic foot per second per square mile (average for 24 hours), and in 1883, which was one of the driest years known for a long period, the flow was .25 of a cubic foot per second per square mile, which is as low a discharge as the large river has ever given. At the same time the small streams were practically dried up, showing that the Merrimac was drawing from the underground sources of supply.

In regard to the tables to which Mr. Hastings has referred, the yields might show a little change if they were arranged according to the order of dryness in each year. On the Merrimac the period from June to August is not always the driest period, but September and October are often the driest months of the year, and November is likely to be one of the wet months. In making comparisons we find it convenient to take a series of years, arranging the months from the driest to the wettest, and taking the averages of the driest months in each year, the second driest, and so on. But Mr. Hastings' arrangements afford the means of very interesting comparisons, and are certainly of great value. It would be of great interest if more records could be kept in regard to the flow of small streams, and, where there are pumping stations and dams, records of what is pumped and what runs to waste would indicate the yield and the run-off of the streams. It is a matter which can be arranged for very readily where there are dams and overflows. Records of this sort are very valuable, and Mr. Newell, the Chief Engineer of the Hydrographic Survey, is making a very important compilation of such records. The data are of great benefit in increasing our general knowledge of the flow of streams.

BY FREEMAN C. COFFIN, CIVIL AND HYDRAULIC ENGINEER,
BOSTON, MASS.

[Read January 13, 1904.]

Engineers and water-works officials are, I think, becoming more and more impressed with the belief that the general use of meters is desirable, and that the method of collecting revenue by means of fixture rates, which is now almost universal, will some time become obsolete and water meters become as universal as gas meters.

There is at present some popular prejudice against their use, people believing that with them they will be restricted in the use of water, or that their bills will be larger.

I do not propose to offer any arguments for the use of meters at this time. We have had some such, notably an excellent one in the recent paper of Mr. J. Herbert Shedd. The results in every city or town which has adopted meters to a considerable extent are substantial arguments for their use.

What I wish to do now is to point out the fact that, viewed as a scientific method of assessing water rates, the meter system, in so far as a basis of rates is concerned, is in almost as chaotic a state as the fixture rate system itself. Several years ago in a paper which the present writer read before this Association, he made the following remarks:

"If the rates in any system are not satisfactory, being either too high or too low, the problem of their adjustment to the requirements of expenditure is one of which it may not be too much to say that no solution has been advanced that considers the conditions of the problem.

"The same may be said of fixing rates for new works. The only method that the writer is cognizant of is that of comparison, and comparison of rates only and not of conditions as well. To collect the schedules of several systems and construct rates from these without regard to such factors as cost of construction per capita, consumption of water per capita, cost of water per million

or thousand gallons, etc., is not a rational method, and gives no assurance that the revenue will be approximately equal to the requirements; yet many if not all rates are fixed in this way. "There can hardly be a theory advanced for the adjustment of fixture rates. It is possible that with the sale of water oy measurement, rates could be devised that would bear some relation to the revenue required. This would, however, be difficult and uncertain until more data upon the subject are available."

These remarks are as true to-day as they were then. When meters are to be introduced to a large extent in any water-works system, some system of charges must be devised.

The object should be to so arrange these charges as to secure just about enough revenue each year to meet all of the expenses of maintenance that are not met from sources other than water consumers, so that the minimum charge shall allow the class paying the lowest rate under meter regulations an abundance of water for reasonable use, but none for waste; also arranged so that people with large houses and a great number of fixtures cannot by exceedingly close calculation escape with paying the equivalent of one faucet rate, and above all to encourage the prevention of leaking fixtures and the accompanying excessive per capita consumption of water.

If water-works officials wish to arrange suitable meter rates, what have they to guide them, where can they look for information? If they ask their consulting engineer for advice, what can he answer them? where can he go for data on the following points?

Namely: How much shall be the charge per thousand gallons? What shall be the minimum annual charge? To what class of service shall it apply? How much water shall be allowed for the minimum rate? If the minimum rate is adjusted for the singlefaucet class of consumers, will the revenue from large houses with many fixtures be commensurate with the comfort and luxury of the service furnished? If not, how can this case be met? How can meter rates be arranged in a place with a large summer and small permanent population?

The above are all practical questions. Who can answer them with any degree of confidence?

The rates must be arranged, however, so that on the one hand

the revenue shall be sufficient, and on the other that the charges to careful consumers shall be equitable and not more burdensome than the old fixture charges.

There must be an accumulation of valuable data upon all or many of these points in the water offices of those municipalities which have a considerable proportion of their services metered, but no one engineer or no one board of water commissioners can collect and digest them.

In view of the foregoing, and in order to bring this matter up for discussion, I make the following motion. It is not pressed for immediate action, but it is hoped that before long this Association, following its past record in undertaking the consideration of problems that call for co-operative action, will carry out the suggestion in some form.

Moved: That a committee of this Association be appointed to consider the question of meter rates for water service, collect and digest information, and submit to the Association a schedule or schedules designed to meet various conditions of a metered service, or a basis upon which such schedules can be arranged; that this committee consist of five members; that the committee be authorized to expend a sum not exceeding fifty dollars in printing, postage, and other necessary expenses involved in the collec

tion of data.