(2) A portion of the outlet sewer being masked, and the overflow coming into operation with the aperture under a small static head.-In this case the sewer is masked to, and the sill is placed at, the same level as in case (1), giving an effective aperture of 1.32 square foot. For this type, when the overflow is in action, the figures will be :

In these calculations it has been assumed that no additional velocity is developed through the aperture by the standing wave caused by the impact against the obstruction. From experiments in the Charlotte Road overflow-chamber, the extra height of such a wave was found to be about 1 inch for every foot per second in velocity. Uncertainty as to the effectiveness of the additional head led to its exclusion, but that it has some value is obvious, and the percentages are therefore slightly less than would actually obtain.

(3) Separating-plate overflow.-By fixing a separating plate, similar to that in the Moseley Street chamber, and with careful calculation, the excess of discharge down the foul sewer practically disappears, as the decreased hydraulic radius under the plate counterbalances the extra velocity of approach of the storm-volume.

The advantages of type (3) over types (1) and (2) are apparent. Accurate separation is rendered practicable, and takes place in the direction of the flow; while the silting due to the sudden checking and back-pounding of the water in type (2) is entirely avoided.

The leap-weir type of overflow is not referred to, as it cannot be successfully adapted to the purpose with any degree of exactitude. The importance of constructing partial bell-mouths, or bonnets, in the outlets from manholes in storm-water culverts, which are usually calculated to flow full, or with a hydraulic gradient, is often entirely ignored; the benching, if any, should also be built level with the crown of the outlet.

When the velocity in the main exceeds that in the affluent, great care must be exercised that the connection is made above the highest floodwater level of the main, as the additional head required to develop the extra velocity in the affluent may lead to serious consequences. The following case is cited to illustrate the importance of this point. A 12-inch affluent-pipe, with a gradient of 1 in 120, discharged into a 2-foot 3-inch by 1-foot 6-inch egg-shaped combined sewer, having an

inclination of 1 in 20, 6 inches above the invert. The flow in the main sewer had never been known to be greater than 18 inches in depth, yet the affluent, although adequate in capacity with a gradient of 1 in 120, was continually back-pounding through its basement connections. The velocity in the main, when flowing two-thirds full was 16.84 feet per second, and that of the submerged branch 4·5 feet per second when full. To develop the extra velocity needed in the affluent in order to allow it to discharge freely into the main required 4 feet additional head, and as this head was not available without flooding basements, the effective gradient was reduced from 1 in 120 to 1 in 216, and the discharge by 25 per cent. Had the connection been made 12 inches higher, or 18 inches above the invert, the gradient would have been reduced only to 1 in 135, and the discharge by 6 per cent.

The foregoing investigations, being limited in scope to the means of an individual, are neither as extensive nor as complete as the Author would wish. Very little money is available for preliminary researches for public works, in comparison with the large sums sometimes wasted through insufficient inquiry into local conditions.

The Author's acknowledgments and thanks are due to Mr. John Price, M. Inst. C.E., City Engineer and Surveyor of Birmingham, for his kindly encouragement, and for permission to conduct the experiments; to the engineers who so courteously supplied valuable information bearing on the subject; and to his colleagues Mr. C. W. L. Alexander, Assoc. M. Inst. C.E., for help with the Pitôt gauge, Mr. Leslie Roseveare, Assoc. M. Inst. C.E., for his adaptation of the figures relating to Plymouth, and Mr. Nugent Weston, Assoc. M. Inst. C.E., and Mr. E. E. W. Butt, for their assistance in the preparation of the diagrams.

The Paper is accompanied by eleven drawings, two sunprints, and one tracing, from which Plate 1 and the Figures in the text have been prepared; also by nine photographs and the following Appendix.

[APPENDIX.

APPENDIX.

GAUGINGS

TABLE II.

OF VELOCITY OF FLOW IN 4-FOOT 6-INCH BY 2-Foor 9-INCH "OLD BIRMINGHAM" EGG-SHAPED BRICK SEWER IN DIGBETH, BIRMINGHAM, CONSTRUCTED IN 1846 TAKEN ON THE 7TH JUNE, 1904.

Length of section gauged = 230 feet.

=

A, area of flow 0.28 square foot; P, wetted perimeter = 1.60 foot; R, hydraulic mean radius 0.175 foot; S, inclination = 1 in 38.7. Calculated velocity, V, by Crimp's formula (V=124 R9 67 S05) 6.23 feet per second.

Velocity as taken by floats :

=

Relation of gauged velocity to calculated velocity=98 per cent.

EXPERIMENTS TO ASCERTAIN THE RELATION BETWEEN DEPTHS OF FLOW OUTSIDE AND INSIDE THE 7-INCH TUBE USED IN CONNECTION WITH THE BALL FLOAT OF AUTOMATIC RECORDER.

TABLE III.-ESTIMATE OF MINIMUM TIME OF FLOW THROUGH THE LONGEST LINE