722LETTERS TO in light nuclei; 6 if we take (2+), we have £2 radiation with (2/+l)r 7 = 250 or 1250 ev and (27+1) \M| 2 = 135 or 675, which seems improbable. (1-) is therefore the most plausible assignment ; it is made almost certain by the angular distribution of the alpha-particles (A. V. Cohen and A. P. French, private communication). If we are correct-in the identification and assignment, we have a breakdown of the isotopic spin and charge parity rules; the emission of alpha-particles demands here T=0, even charge parity, while the emission of El radiation demands here T=l, odd charge parity; 1 ' 2,7 yet both widths are large (r 7 =150 ev; r a = 75 kev) and neither may be supposed to have suffered a very large measure of discouragement. We may not rule out the possibility that there are two resonances and that what we are observing is the rules in action rather than their violation, though this seems unlikely in view of the agreement in position and width. The excitation in O 16 is 13.1 Mev and the first T= 1 level may be expected at about 12.5 Mev. 8 The reaction B n (p, «)Be 8 (ground state) is resonant at somewhat over 1 Mev 9 in proton energy; B n (p, y)C 12 (ground state) is strongly resonant at 1.4 Mev 10 with a large radiative width. This may be a similar example but is not so clear-cut. The reactions Be 9 (&a)Li 6 and Be 9 (p,d)Be 8 are resonant at 0.94 Mev, 11 while Be 9 (^>, Y)B 10 has a strong, almost certain, El resonance of similar width at 0.998 kev; 12 but these may well involve two different states. In these last two examples we are in a region of excitation containing T=l states. W E have carried out two experiments to attempt to assess the purity of isotopic spin or charge parity states 1 " 3 of moderate excitation.O 16 possesses 4 a (1-) state at 7.12 Mev, a (2+) state at 6.91 Mev, a (3-) state at 6.14 Mev and a (0+) state at 6.05 Mev; the ground state is (0+). All these states are expected 5 to have r=0; if we think in terms of charge symmetry alone, the charge parity is probably even. The (1 -) state decays to the ground state thereby violating the isotopic spin 6 or charge parity 7 rule. The E2 decay to the (3-) state is uninhibited by the special rules; we have shown it to occur at least 120 times less probably than the forbidden El transition. We have also shown that the (2+) state decays to the ground state at least 200 times more readily than to the (3-) state, although, in the absence of the special rules this latter El transition would be preferred. The singleparticle matrix elements 8 seem to be unexpectedly reliable 9 for the prediction of El radiative widths, and there is no evidence that they are grossly wrong for E2 transitions in light nuclei; if we apply them to this case, we obtain the result that the contamination of the (1 -) state is more than 0.2 percent in amplitude and that that of the (2+) and (3 -) states is less than 3 percent in amplitude (assuming the ground state to be pure r=0). These estimates are probably reliable to a factor of five or better.The first sta...