100 RATIONAL CONSTRUCTION OF FURNACES sufficient to show clearly the reasons which have led to the com- plete abandonment of the Withwell stove. ^ The ideas of Withwell have found their realization in the Massick and Crook stove, which had quite a name at one time, and which was only distinguished from the Withwell stove by the arrangement of the passes. This arrangement was such that the gases traveled alternately up and down through the three passes. The Massick and Crook hot-blast stove has now been superseded for reasons which will be clearly evident to the reader. Many inventors have applied themselves to the problem of FIG. 62. FIG. 63. securing uniformity in the operation of the checkerwork heating surface and passes, apparently without the slightest suspicion that the regular and uniform descent of the hot gases in cooling is a natural property of the gases. For a time the design patented by Becker (Fig. 63) was much favored. In this design the open- ings in the checkerwork were given different dimensions, naturally a> Withwell himself noted the defects in his design, and in a later design (Fig. 62) he gave the hot gases a circulation in the proper direction, from top to bottom. This design approximates the type of regenerator shown in Fig. 51.