120 SECTIONAL ADDRESSES. 
2. The substitution, where feasible, of rockfill dams for those 
of masonry or monolithic concrete. 
3. The introduction of outdoor installations with the minimum 
of power-house construction. 
4. The simplification of the power plant. 
Some progress has already been made in these directions, and it is 
probable that experience based on recent installations and experimental 
investigations will enable considerable further progress to be made. 
Research in Hydro-Electric Problems.—There are few branches of 
engineering in which research is more urgently required and in which 
it might be more directly useful. 
Among the many questions still requiring investigation on the 
civil and mechanical side may be mentioned— 
1. Turbines.—Investigation of turbine corrosion as affected by 
the material and shape of the vanes. 
Effect of erosion due to sand and silt. 
Resistance to erosion offered by different materials and coatings. 
Bucket design in low head high-speed turbines. 
Draft tube design. 
Investigation of the directions and velocities of flow in modern 
types of high-speed turbines. 
Investigation of the degree of guidance as affected by the number 
of guide and runner vanes, 
2. Conduits and Pressure Tunnels.—The design of large pipe 
lines under low heads with the view of reducing the weight 
of metal. The investigation of anti-corrosive coatings, so 
as to reduce the necessity for additional wall thickness to 
allow for corrosion. 
Methods of strengthening large thin-walled’ pipes against bend- 
ing and against external pressures. 
Methods of lining open canals and of boring and lining pressure 
tunnels. 
Effects of curvature in a canal or tunnel. 
3. Dams.—Most efficient methods of construction and_ best 
form of section especially for rockfill and earthen dams. 
Best methods of producing water tightness. 
4. Run-off data.—Since the possibility of designing an instal- 
lation to develop the available power efficiently and 
economically depends in many cases essentially on the 
accuracy of the run-off data available, the possession of 
accurate data extending over a long series of years is of great 
value. 
While such data may be obtained either from stream gaugings or 
from rainfall and evaporation records, the former method is by far the 
more reliable. For a reasonable degree of accuracy, however, records 
must be available extending over a long period of years, and at the 
present moment such data are available only in very few cases. 
Where accurate rainfall and evaporation records are available, it is 
