GENERAL DISCUSSION Prof. J. J. van Loef and Dr. A. M. van der Kraan (Reactor Institute Delft) said It is suggested in the paper by Lyubutin et al. that the observation of the quadrupole interaction at the 57Fe nuclei in polycrystalline YIG below T is practically impossible. In analyzing our Mossbauer spectra of polycrystalline YIG at room temperature shown in fig. 1 we come to the following conclusions.1 The measured magnetic FIG. 1 .-The Mossbauer spectrum of polycrystalline YIG as an absorber and 57C0 in Pd as a source. The arrows indicate the line positions of the various hyperfine components. hyperfine spectrum of YIG comprises three different Zeeman splittings two for Fe3+ ions in the a-sublattice and one for Fe3+ ions in the d-sublattice. The two a-site hyperfine spectra are characterized by two quadrupole interactions of +e2qQ = + (0.05&0*015) mm/sec for 75 % of the Fe3+ ions and -(0-12f0.03) mm/sec for 25 % of the Fe3+ ions respectively.Furthermore the hyperfine fields at these two groups of Fe3+ ions are different by (6.2+ 1*7)kOe which is in the first place due to the inequivalence of the octahedral iron sites because of dipolar fields. This is in good agreement with the results of n.m.r. experiments by Meyer and co-workers,2 who found a difference of (6.6+0.3)kOe. The d-site hyperfine spectrum is a single six-line Zeeman splitting with no quadrupole interaction in line with the results reported by Lyubutin. We have compared the Mossbauer spectra obtained with polycrystalline YIG with that found in a single crystal without applying a magnetic field.Both the single crystal and the powdered sample give the same results. Therefore it is con- cluded that even in polycrystalline samples of the garnet the electric hyperfine inter- action measured at temperatures T<T can be fully understood. In fact from the correlation of the quadrupole coupling constant with the magnetization vector the easy direction of magnetization follows in a straightforward manner. We have applied this method to determine the easy direction in Sm1G.l The major reason why our YIG results deviate from those reported by Lyubutin and co-authors is probably because our counting rates per velocity channel of about lo6 are by an order of magnitude higher than theirs. J. J. van Loef Proc. Int. Con$ Magnetism (Boston 1967). R. Gonano E. Hunt H. Meyer and A. B. Harris J.Appl. Physics 1966 37 1322 and private communication. 38 GENERAL DISCUSSION 39 Prof. N. N. Greenwood (Newcastle upon Tyne) said Is it possible to apply Gilleo's theory to measurements taken only at 77"K? Lyubutin states that dependence of the Curie temperature on x cannot explain the results by itself but surely this dependence must play at least some part in the analysis. It also seems strange that the isomer shifts and quadrupole interactions for both sites of Fe3+ ions are almost independent of the tin content of the garnets. Is there any explanation of this? In particular the electric field gradient would be expected to be sensitively dependent on lattice distortions brought about by the varying tin content. Dr. R. Krishnan (Laboratoire de Magnitisme France) said With regard to Lyubutin et aZ.'s paper when one interprets the effect of substitution of non-magnetic ions on the hyperfine fields of garnet and spinel ferrites in terms of the variation in the number of magnetic neighbours of Fe3+ in the two sites one is faced with cer- tain difficulties.Fur example in the present case of Y3-,Ca Fe,-,Sn,0,2 the hyperfine fields at 77°K for both a and d sites decrease as x increases whereas in many cases such as Nil-,Zn,Fe204 studied by Goldanski et aZ. and Nil+ GeXFe2-2,O4 studied by Fatseas and Krishnan the hyperfine field at only one site varies. In the latter system we have shown that the hyperfine field at A sites decreases as x increases whereas that at B sites remains constant. We had attributed this to the increased degree of covalency in Fe3+-02- bonds for A sites by virtue of the presence of Ge4+ ions in these sites. Thus we see that the nature of chemical bonds is more important than just the number of magnetic linkages (Gillea model). T. Fatseas and R. Krishnan Int. Con$ Magnetism Boston Sept. 1967.