by steel-trestle approaches entirely disconnected from the main structure.

Fourth. Dr. Steinman uses a tension intensity of working stress of 20,000 pounds for carbon steel and one of 30,000 pounds for nickel steel,

while the author's practice has been to employ, respectively, 16,000 pounds and 28,000 pounds.

Fifth. Dr. Steinman ignores entirely the effect of impact on trusses, while the author allows for it. In the very long spans this cuts but little figure; however, such is not the case for the shorter spans.