In another part of this job (the factory annex) where, owing to the open nature of the structure, it was impossible to house it in as well as the warehouse which had bearing walls to curtain off the sides, less fortunate results were obtained.  A temperature drop over night of nearly 50 deg., followed by a spell of alternate freezing and thawing, effected the ruin of at least the upper 2 in. of a 6-in. slab spanning 12 ft. (which was reinforced with 1/2-in. round bars, 4 in. on centers), and the remaining 4 in. was by no means of the best quality.  It was thought that this particular bay would have to be replaced.  Before deciding, however, a test was arranged, supports being provided underneath to prevent absolute failure.  But as the load was piled up, to the extent of nearly 400 lb. per sq. ft., there was no sign of giving (over this span) other than an insignificant deflection of less than 1/4 in., which disappeared on removing the load.  This slab still performs its share of the duty, without visible defect, hence it must be safe.  The question naturally arises:  if 4 in. of inferior concrete could make this showing, what must have been the value of the 6 in. of good concrete in the other slabs?  The reinforcing in the slab, it should be stated, was continuous over several supports, was proportioned for (w l)/8 for the clear span (about 11 ft.), and three-fourths of it was raised over the supports.  This shows the value of the continuous method of reinforcing, and the enormous excess of strength in concrete structures, as proportioned by existing methods, when the reverse stresses are provided for fully and properly, though building codes may make no concession therefor.

Another point may be raised, although the author has not mentioned it, namely, the absurdity of the stresses commonly considered as occurring in tensile steel, 16,000 lb. per sq. in. for medium steel being used almost everywhere, while some zealots, using steel with a high elastic limit, are advocating stresses up to 22,000 lb. and more; even the National Association of Cement Users has adopted a report of the Committee on Reinforced Concrete, which includes a clause recommending the use of 20,000 lb. on high steel.  As theory indicates, and as F.E.  Turneaure, Assoc.  M. Am.  Soc.  C. E., of the University of Wisconsin, has proven by experiment, failure of the concrete encircling the steel under tension occurs when the stress in the steel is about 5,000 lb. per sq. in.  It is evident, therefore, that if a stress of even 16,000 lb. were actually developed, not to speak of 20,000 lb. or more, the concrete would be so replete with minute cracks on the tension side as to expose the embedded metal in innumerable places.  Such cracks do not occur in work because, under ordinary working loads, the concrete is able to carry the load so well, by arch and dome action, as to require very little assistance from the steel, which, consequently, is never stressed to a point where cracking of the concrete will be induced.  This being the case, why not recognize it, modify methods of design, and not go on assuming stresses which have no real existence?