evident underlying principle involved in this statement; but the published records of some of the greatest bridges designed by English engineers show the use of pieces of trusses so curved that actually there is compression on one extreme fibre and tension on the other. Architectural effect is undoubtedly a very commendable feature in bridge-designing; but its adoption is no excuse for the violation of the fundamental principle that every compression or tension member of a truss or open-webbed girder should be absolutely straight from end to end.

It seems almost unnecessary to state that the appearance of curvature can be obtained by employing short panels and making each chord-length straight between panel points.

Principle XVII.

In all structural metal-work, excepting only the machinery for operating movable parts, no torsion on any member should be allowed if it can possibly be avoided; otherwise, the greatest care must be taken to provide ample strength and rigidity for every portion of the structure affected by such torsion.

It is not often that this question arises; nevertheless it is sometimes forced upon the consideration of the engineer. It came up lately in the author's practice in the case where an elevated-railroad exit-stairway, having at mid-height a landing and a 180-degree turn, had to be supported by a single column in order to comply with the demands of adjacent property owners.

Principle XVIII.

The gravity axes of all the main members of trusses and lateral systems coming together at any apex of a truss or girder should intersect in a point whenever such an arrangement is practicable; otherwise the greatest care must be employed to insure that all the induced stresses and bending moments caused by the eccentricity be properly provided for.

 This is an important rule that is more often honored in the breach than in the observance;  in  fact,  as far as the author