Ahokas et al.[1] measured the hyperfine frequency shifts in three-dimensional spin-polarized atomic hydrogen by means of ESR. In this Comment, we address their analysis of the interaction energy of the ground-state H atoms in different hyperfine states and show that the quoted difference Áa ¼ a t À a s between the triplet and singlet scattering lengths derived from the correctly measured shifts is overestimated by a factor of 2.Ahokas et al. observed the transitions a ! d and b ! c in the presence of the third-state atoms (b and a, respectively) and found the shifts of the corresponding resonance fields to be ÁB ad ¼ C ab n b þ C aa n a and ÁB bc ¼ C bb n b þ C ba n a with C ab % C ba ¼ 8ð2Þ Â 10 À19 cm 3 and C aa ; C bb ( C ba . To explain this observation, Ahokas et al. expressed the spin states of a pair of atoms in the jS; m S ; I; m I i basis, that is, in terms of electron and nuclear singlets and triplets. In particular, in the high-field limit, jabi ¼ 1 ffiffi 2 p ðje t ; n t i þ je t ; n s iÞ, jaci ¼ jbdi ¼ 1 2 ðje t ;n t i þ je t ;n s i þ je s ;n t i þ je s ;n s iÞ. At low temperature, only the symmetric components je t ; n t i and je s ; n s i contribute to (s-wave) scattering via the triplet V t and singlet V s interatomic potential, respectively.More specifically, the interaction energy of H atoms in different hyperfine states and (; ¼ a; b; c; d) is given by the second-quantization Hamiltonian (i; j ¼The wave function of two bosons must be symmetric,where the spatial and spin parts are, respectively, R AE ¼ 1 ffiffi 2 p ½c k ðr 1 Þc q ðr 2 Þ AE c q ðr 1 Þc k ðr 2 Þ and jiji AE ¼ 1 ffiffi 2 p ðjiji AE jjiiÞ. The use of the symmetric form (2) of the diatomic wave function instead of simply jki; qji ¼ c k ðr 1 Þc q ðr 2 Þjiji does not change the sum (1) because the bosonic creation (annihilation) operatorsâ þ ki andâ þ qj (â ki andâ qj ) with i Þ j obviously commute. The interaction strength of the pseudopotential Vðr 2 À r 1 Þ ¼ ðr 2 À r 1 Þ has the eigenvalues s or t corresponding to the spin states je s i or je t i of the atomic pair. As such, the potential is nearly independent of nuclear spins: he t n s jje t n s i ¼ he t n t jje t n t i ¼ t and he s n t jje s n t i ¼ he s n s jje s n s i ¼ s (here we write only the spin parts of the matrix elements). Instead, Ahokas et al. used he t n s jje t n s i ¼ he s n t jje s n t i ¼ 0, arguing that the antisymmetric states do not scatter via s waves. This zeroing is only valid if the matrix elements include the vanishing spatial factor hR À jðr 2 À r 1 ÞjR À i. Actually, it is this spatial factor which cancels the contribution of the antisymmetric states to s-wave scattering and to the interaction energy (1). On the contrary, the spatial part of the matrix elements for the symmetric states R þ jiji þ is doubled. In other words, the atoms of a heterostate symmetric pair behave as identical. Clearly, jabi þ ¼ je t n t;0 i and jaci þ ¼ jbdi þ ¼ 1 ffiffi 2 p ðje t;0 n t;0 i þ je s n s iÞ. Consequently, þ ab habjjabi þ ¼ t and þ ac ¼ þ bd ¼ 1 2 ð s þ t Þ. Then, ...