THE question of how great a gain in depth dose is obtained by the use of kilovoltages above 200 is often answered merely by stating the values for a depth of 10 cm. (D10), obtained for various fields by using a thimble chamber in a large phantom, and comparing them to those obtained under similar conditions with a 200 kv. apparatus. For example, in the case of a field 10 × 10 cm. the D10 for radiation from our 1,000 kv. tube is more than 30 per cent greater than that from our 200 kv. tube, when a field of the same size is used with the same target to skin distance. This statement is impressive, for it indicates that if the same skin dose is maintained, more than 30 per cent greater dose can be delivered to a deep tumor. Further, if certain practical factors were left out of consideration, a gain of more than 60 per cent in the dose to the tumor might be anticipated by the use of two fields and, by using multiple fields, an enormous gain might be expected. In actual clinical practice, however, the relative advantages of radiations of two different qualities are not readily pictured by stating a few comparative numbers, but are dependent, in a variety of ways, on the clinical problems involved. The general necessity for using cross-fire technic in treating most deep neoplasms with externally applied x-rays or gamma rays, makes necessary the comparison of surface as well as depth doses under different conditions. With these points in mind, we compared the results of cross-firing beams through thicknesses of from 12 to 24 cm., using data obtained both in “unlimited” phantoms and in those limited to the specified values. By an “unlimited” phantom is meant one in which there is at least 15 em. of material below each level to provide almost the maximum possible amount of back-scattered radiation. It is inconvenient to use water for limited sections because of the difficulties involved. Wax has certain unfavorable properties (1). At the Memorial Hospital, in New York City, a cellulose material, known commercially as Masonite Presdwood (untempered), was found to be a medium very comparable to tissue for use as a phantom. This is obtainable in sheets from about 3 to 12 mm. thick which can be very conveniently stacked to make phantoms of the desired thickness. We used this material in the present study. The ionization measurements were made with a standard Victoreen condenser r-meter using a 25 r thimble chamber. The constancy of output of the apparatus was under continual measurement, making the mutual consistency of the data accurate (less than 1 per cent of error). All surface values refer to readings with the chamber one-half immersed in the surface.