Modern genetic analysis and manipulation of soybean ( Glycine max) depend heavily on an efficient and dependable transformation process, especially in public genotypes from which expressed sequence tag (EST), bacterial artificial chromosome and microarray data have been derived. Williams 82 is the subject of EST and functional genomics analyses. However, it has not previously been transformed successfully using either somatic embryogenesis-based or cotyledonary-node transformation methods, the two predominant soybean transformation systems. An advance has recently been made in using antioxidants to enhance Agrobacterium infection of soybean. Nonetheless, an undesirable effect of using these antioxidants is the compromised recovery of transgenic soybean when combined with the use of the herbicide glufosinate as a selective agent. Therefore, we optimized both Agrobacterium infection and glufosinate selection in the presence of L-cysteine for Williams 82. We have recovered transgenic lines of this genotype with an enhanced transformation efficiency using this herbicide selection system.

Satpathy, Zhang, and Salehpour Reply:The Comment of Ho, Chan, and Soukoulis 1 (HCS) contains three points: (1) the question concerning how good the scalarwave approximation is; (2) misplacement of the experimental point in Fig. 4(b) in Ref. 2;and (3) that the scalar-wave equations can be solved by a matrix diagonalization scheme.Concerning the first point, we agree with HCS that the vector nature of the electromagnetic (EM) waves could indeed be important. This is what we said in the last sentence of our original paper and what motivated us to perform the full vector-wave calculation for the periodic structure. 3 We would like to point out that, contrary to the implications of HCS, the single-sphere Mie resonances are unimportant for the experimental dielectric structure. 4 HCS find that the single-sphere Mie resonances of the isolated sphere for the vector EM waves occur at 12.04 and 15.4 GHz. But actually there are no strong Mie resonances when one calculates the scattering coefficient gsca within the Mie theory 5 for single-sphere scattering corresponding to the experimental parameters (background refractive index /i^ =3.5, sphere refractive index /i fl ""l, m=n a /rib -0.286, sphere filling fraction -0.86, and fee lattice constant =12.7 mm). This is shown in Fig. 1 (case m =0.286) where we see no significant peaks in the scattering coefficient indicating that the Mie resonance effects are small in this specific experimental dielectric structure. Furthermore, even if the Mie scattering effects were large, the single-sphere Mie result may not be expected to carry over into the periodic structure of the experiment where the "spheres" occupy as much as 86% of space.For larger values of the sphere refractive index compared to the background index, however, the scattering cross section does contain strong peaks (Fig. 1, case m = 1.5, for instance). In this case, the Mie resonance effect may show up in the photon bands if the sphere packing fraction is not too large. 3 Concerning the second point of HCS, the experimental point was indeed misplaced in Fig. 4(b) and it has now been corrected in the Erratum. 2 No specific conclusion was really drawn in Ref. 2 from the position of this single data point. As far as comparison with experiments pertaining to the EM waves is concerned, the new vector calculations 3,6 supercede the results of the scalar calculations.Concerning the last point of HCS, it is indeed true that the scalar-wave problem is reducible to a matrix diagonalization scheme. Of course, all results remain unchanged whether one uses a root-search scheme or a matrix diagonalization scheme. The results of the scalarwave calculations of Leung and Liu 7 using the matrix diagonalization scheme agree with the results of our original paper.We thank K. M. Leung and E. Yablonovitch for fruit-