With reference to the matter of formula for steel columns, dealt with in recent articles and letters appearing in Engineering News-Record, the writer wishes to take this opportunity of drawing attention to the fact that existing column formulae are not founded on really proper tests, and that, therefore, engineers should not rest satisfied until such proper tests have been made, and, from the empirical rules deduced therefrom, a really sound column formula has been evolved.

No column experiments have yet been made which conform to actual conditions existing in modern bridges, as very few columns nowadays have hinged ends, and flat- end tests are likely to be very deceptive, because of corner bearing.

The opinion expressed by Mr. Charles E. Fowler in his letter on page 343 of your issue of Feb. 13, 1919, to the effect that a straight-line formula is to be desired on account of its simplicity, is to be commended, as there is every probability that a satisfactory straight-line formula can be derived from proper tests. It is to be noted that some of the formulae given by Mr. Fowler in his letter use the (l/r)², this being a holdover from the old Euler formula, which was established for greater values of l/r than are used in practice.

With reference to Dr. Waddell's scheme for testing top-chord members by means of experiments conducted on a specially-built, full-size bridge, the writer begs to make the following suggestions for a practical method for making such tests: Build a single-track, Class 70, through-truss bridge, with the weak member designed for, say, Class 40 loading, so that the destruction of the weak member could be accomplished without injury to the rest of the structure, thus facilitating repairs for repeated tests. The connection plates for the weak members should be extra strong. After the compression tests are completed, a series of corresponding tests on tension members should be made, so as to determine the real efficiencies of both tension and compression members and their comparative strengths. Unless this kind of a test be carried out, we shall never have any really reliable test data, as the existing data are, to a great extent, misleading. It is advisable that more than one value of l/r should he tested for.

It should not be overlooked, in making up a design for a test structure, that there is a necessity for thorough detailing throughout, especially in connection with the weak members, so that failure will be in the form of a square break instead of being due to any insufficient detail, thus removing ambiguity. It would be advisable to build the permanent members of the trusses of the test bridge of nickel steel, so that the permanent and the weak sections would be somewhat alike, if the weak section were of the ordinary structural grade of carbon steel.

The most satisfactory method of applying the load would be by means of hydraulic presses or jacks, to be applied at panel points, and all connected so that the load would be uniform. Pressure records should be automatically made. Control of the applied load by this means would be very simple and positive. It would be desirable to make tests for low values of l/r, so as to establish the necessity or otherwise for limiting intensities of working stress.

Many of the tests we are relying on at the present time for our formulae were made on badly designed struts; and I venture to say that tests on properly designed struts would show up stronger than the present tendency of the profession in strut design indicates.

If the series of tests above suggested were thoroughly carried out, the engineering profession would possess some authentic knowledge concerning the actual strengths of both tension and compression members in bridges, and a large amount of guess work would be avoided in future constructions.

As yet no steps have been taken to materialize this suggestion of the author's. Probably the profession will say that it is up to him to see that his proposed tests are made; and, possibly, after this treatise is completed and launched,  he  will  find  time  to  make  an  attempt to induce the