BUREAU OF PUBLIC ROADS. 477 



B}' merely rounding the entrance to a 24-incli vitrified clay pipe 

 culvert the capacity may be increased approximately 13 per cent over 

 that obtained in a square entrance. By increasing the area of the 

 cross-section of the outlet end of an 18-inch vitrified clay pipe cul- 

 vert so that the area is about double and the angle of divergence 

 about 10°, the discharge of the culvert, ^vhen the outlet is submerged, 

 mav be increased 40 per cent over that obtained in the same culvert 

 having a uniform bore throughout. 



The utilization of the data obtained in culvert design enables 

 an engineer to increase the carrying capacity of culverts 40 to 50 per 

 cent over former standard practice with a comparatively small in- 

 crease in cost. It is proposed to continue the investigations to 

 include box culverts. 



At the laboratory at St. Paul, Minn., maintained jointly by the 

 bureau, the University of Minnesota, and the Minnesota Depart- 

 ment of Drainage and Waters, the investigation of the problem of 

 the disintegration of concrete tile by alkali was continuecl. Experi- 

 ments were also made on the effect of freezing on clay tile. One 

 of the important phases of the work at this laboratory has been 

 the working out of a method of measuring the progressive de- 

 crease in crushing strength of specimens of concrete when sub- 

 jected to alkali conditions simulating as closely as practicable those 

 of ordinary bad field conditions with respect to alkali. 



It was found after extensive tests that the increase in volume of 

 a specimen — this increase being determined by accurately measur- 

 ing the increase in length — bears a relation to the decrease in crush- 

 ing strength. By this means it is possible within limits to deter- 

 mine at any time the strength of a specimen without testing it to 

 destruction. A series of tests has been conducted to determine the 

 influence of consistencj', mixture, and curing conditions on re- 

 sistance to disintegration. It is yet too early to show the complete 

 results of these experiments, though in general they indicate that 

 consistency is the least important of the three variables mentioned. 

 The amount of cement in the mix and the curing conditions evi- 

 dently have a far-reaching effect upon the life of the specimen. It 

 appears evident, however, from the work to date that the influence 

 of both these factors on the resistance to disintegration is quite 

 limited. Some 1,800 samples of concrete drain tile in use in the 

 alkali-infested region have been tested with a threefold object in 

 view: first, to ascertain the quality of tile in general use in the 

 State; second, to determine how concrete tile so tested compared 

 with laboratory-made specimens; third, to promote a general im- 

 provement in the quality of drain tile used in the State. The data 

 on the practical phases of the manufacture of tile have not been 

 complete^ analyzed, but it is not too early to state that a funda- 

 mental fault in the present method of manufacturing small-sized 

 concrete tile in the ordinary packerheacl machine is the very dry 

 mix necessary in order that the jackets may be easily stripped. 



One of the first essentials for better concrete tile in the smaller 

 sizes is better manufacturing method. This fact is realized by manj^ 

 of the manufacturers of concrete tile in Minnesota and elsewhere 

 who have cooperated willingly in the work. Wlien the work was 

 started very little concrete tile was being manufactured that would 

 pass the requirements of the American Society for Testing Materials 



