fits of the folding model with no density dependence. Using the obvious relation Ar m 2 =Zr p 2 + Nr n 2 and the value 1 r^Ca) =3.386 fm one obtains r"( 48 Ca) -r,( 48 Ca) =0.17± 0.10 fm. Table II compares the present results with previous ones, all of which were obtained using simple analytic functions for the nuclear-matter distribution. If the uncertainties obtained in the present work are typical of the uncertainties that should have been quoted for previous results, then all seemingly conflicting results are in agreement with each other. We thank Professor Dr. G. Schatz and Professor Dr. P. Brix for their encouraging interest and the cyclotron crew for good cooperation. An unambiguous and model-independent separation of direct and compound elastic cross section was performed by a simultaneous fluctuation analysis of the cross section and the analyzing power for the reaction 30 Si(/> pol ,£ 0 ) 30 Si. The compound elastic cross section was calculated by means of Hauser-Feshbach theory without any adjustable parameters. Comparison of experiment with the calculation results in a compound-elastic enhancement factor of 2.09±0.14 for this strong-absorption, many-channel case.In this Letter we report definitive experimental evidence for a factor-of-2 enhancement of the compound elastic cross section (a CE ) over the Hauser-Feshbach 1 cross section (a HF ) in a reaction with many strongly absorbing channels. This so-called compound-elastic enhancement factor W acc =(j CE /a m has been the subject of considerable theoretical efforts. 2 " 6 In the many-channel, strong-absorption limit most theorists now agree that W aa should be two irrespective of the number of direct reaction channels. Previous attempts to determine W aa experimentally have been performed only with unpolarized protons yielding a large model-dependent error. 7 In the best of these experiments Ernst, Harney, and Kotajima 7 have confirmed the effect of channel self-correlation (W cca > 1) but their values of W aa ranged from 1.6 to 3.5. Thus a clear statement about the validity of the underlying compound-nucleus theories was not possible. Since the enhancement factors derived from these theories are widely used, 8 * 9 it is important to have a much more critical test.A model-independent separation of a CE and a DI (DI indicating the direct part of the cross section), necessary for an accurate determination of W«a> can be performed by measuring both the differential cross section a and the analyzing power A (9) either in an energy-averaged experiment 10 or in an Ericson fluctuation experiment 11 on a spin-0 target nucleus. We have used both methods: First we have applied the method of Henneck and Graw, 11 where a combination of the normalized variances of cr and uA yields the relative direct contribution y D to the reaction without any assumption about the effective number of spin channels iV eff 0 We then checked the consistency of this method by describing the extracted cP l with the optical potential obtained by fitting the artifically energy-average...